we won't have a project meeting today with many of the team at the Openstack Summit. We will reconvene next week. If there are any pressing matters bring it up on IRC.
Please continue to complete any outstanding Trello cards. Also feel free to pick up any outstanding bugs.
Thanks, Sam
Canceled Event: OpenDaylight - OPNFV community sync meeting @ Thu May 21, 2015 10am - 11am (dneary@redhat.com)
This is a monthly call to allow integration issues between OpenDaylight and OpenStack in the context of OPNFV to be shared and resolved, and to encourage better communication between OPNFV and the OpenDaylight project.
Agenda this month: * OpenDaylight issues in BGS (Fuel and Foreman participants: Please communicate any current issues you have to allow me to add them to the agenda) * Deploying OpenDaylight in a Docker container (Dan Smith, any issues?) * OpenDaylight development/roadmap process: How can OPNFV participants define priorities and communicate feature gaps to the OpenDaylight community? https://wiki.opnfv.org/community/opendaylight
Dial-in numbers:
US Toll-Free Dial-In Number: 800 451 8679 US local dial-in number: +1 (212) 729-5016
Global Access Numbers Local:
China Domestic Dial-in # 4006205013 China Domestic Dial-in # 8008190132 Finland Helsinki Dial-in # 0923194436 France Paris Dial-in # 0170377140 Germany Berlin Dial-in # 030300190579 Germany Frankfurt Dial-in # 06922222594 Spain Madrid Dial-in # 914146284 Sweden Stockholm Dial-in # 0850513770 United Kingdom Dial-in # 02035746870 United Kingdom LocalCall Dial-in # 08445790678
Global Access Numbers Toll-Free
Australia Dial-in # 1800337169 France Dial-in # 0805632867 Germany Dial-in # 08006647541 India Dial-in # 180030104350 Japan Dial-in # 0120994948 Japan Dial-in # 00531250120 Netherlands Dial-in # 08000222329 Spain Dial-in # 800300524 Sweden Dial-in # 0200896860 Switzerland Dial-in # 0800650077 United Kingdom Dial-in # 08006948057
You are receiving this courtesy email at the account ovsdb-dev@... because you are an attendee of this event.
To stop receiving future updates for this event, decline this event. Alternatively you can sign up for a Google account at https://www.google.com/calendar/ and control your notification settings for your entire calendar.
From: Prateek Garg <prategar@...> Date: Monday, May 4, 2015 at 12:41 PM To: Edward Warnicke <hagbard@...>, "Amit Mandke (ammandke)" <ammandke@...> Cc: "ovsdb-dev@..." <ovsdb-dev@...> Subject: Re: [ovsdb-dev] [Need Help] Arbitrary connection port is part of the node-id
Hi,
I have implemented this change and submitted the same:
In case of ptcp connection, save the iid of the ovsdb node in the external ids of the open_vswitch table when ovs connects to controller. This is done in case of ptcp connect only as tcp would be
using system-id for iid which is not going to change in case of disconnect and reconnect of the same ovs.
In OvsdbConnectionInstance the iid for the ovsdb node would be genereated based on external ids of the ovs. If the iid is not present in the external ids the iid would be generated using the system-id (in case of tcp). There is also a fail safe, for the
case when both system-id and open daylight-iid is not present in external_ids of open-vswitch which is to use the old way of generating the iid for the ovsdb node.
The iid generated in the step 2 is provided to all required commands in OvsdbOperationalCommandAggregator to make sure the iid for ovsdb node is generated just once and all the commands use what they get and not have to regenerate the iid for the ovsdb
node.
The overall idea is to use the system-id for iid of ovsdb node when system-id is present in external-ids of the open-vswitch when connection in tcp mode. This is because the port changes when ovs reconnects to controller in tcp mode. Further, in all other
scenarios the old way would be used.
Regards
Prateek
From: Edward Warnicke <hagbard@...> Date: Friday, April 24, 2015 at 8:50 PM To: "Amit Mandke (ammandke)" <ammandke@...> Cc: Prateek Garg <prategar@...>, "ovsdb-dev@..." <ovsdb-dev@...> Subject: Re: [ovsdb-dev] [Need Help] Arbitrary connection port is part of the node-id
If we do not receive the initial dump of the DB... we aren't getting anything to work at all... it s a fundamental failure.
And always remember that OVSDB can connect to us unsolicted... so we will always have to deal with ovsdb instances that we *didn't* create the connection for. This is another reason that the solution I'm advocating for here is good... it works consistently
in active and passive mode.
Ed
On Fri, Apr 24, 2015 at 7:08 PM, Amit Mandke (ammandke)
<ammandke@...> wrote:
Isn’t that tricky for a person who is trying to connect ovs to controller? Its like you may or may not see the node in operations based whether or not it succeeds or fails to update.However the ovs would show conencted=true and manager ip, with no data
in operations. :-)
IMO connection received is exclusive event for creating ovsdb node and then registering for callbacks for update. Current implementation seems more logical(to me) than changing it for 1 case of IID creation.
My further 2 cents on iid serialization/deserialization:
In general, IMO this IID serialization and deserialization is a risk. And should be avoided if possible. If a system is running for long period of time and for some reason someone connected disconnected same ovs nodes with diff setting once with IID
passed to it once without. Your config would be looking very out of synch of your operations. On top of that if the apps that are trying to use this data programmatically would most likely fail to take care of such situations.
Also different apps may choose different strategies to generate iids and would be impossible for them to coexist.
So having only one way of creating IIDs would help in longer run. If one wants avoid iid generation while connecting to OVS or creating bridges then we can always expose RPCs for that. However, if one wants to use CONFIG store to configure, then they
have to build appropriate iid.
Subject: Re: [ovsdb-dev] [Need Help] Arbitrary connection port is part of the node-id
Not so tricky... we either get that dump or we are borked...
Ed
On Fri, Apr 24, 2015 at 2:07 PM, Amit Mandke (ammandke)
<ammandke@...> wrote:
It sounds tricky, cause once you register for callbacks you can’t be sure of the sequence of all the updates right? Or even guaranty if each update would be processed?
So essentially you are letting go a sure-shot event where you received the connection to store the OVSDB node and relying on the future updates to create it.
Subject: Re: [ovsdb-dev] [Need Help] Arbitrary connection port is part of the node-id
I had an idea last night as I was falling asleep that might be interesting.
How about we create the ovsdb-node in operational with the OpenVSwitchUpdateCommand (which should get tripped with the original dump of data from ovsdb on connect) rather than invoking OvsdbNodeCreatedCommand in the OvsdbConnectionInstance() constructor.
That way you have access to the info you need, don't have to do a select (and thus a new round trip) etc.
Thoughts?
Ed
On Wed, Apr 22, 2015 at 3:40 PM, Prateek Garg (prategar)
<prategar@...> wrote:
This can be done in OvsdbNodeCreateCommand. The issue I see there is changing the signature of the OvsdbNodeCreateCommand’s constructor.
Let me know if its ok to change the signature and pass OvsdbClient.
However, I believe still we would need to run a select statement to get the system-id in OvsdbNodeCreateCommand.
Please let me know your thoughts.
Thanks
Prateek
From: Edward Warnicke <hagbard@...> Date: Wednesday, April 22, 2015 at 3:26 PM
And why change ConnectionInfo (that really really is info about the ip and port...)
Ed
On Wed, Apr 22, 2015 at 3:25 PM, Edward Warnicke
<hagbard@...> wrote:
Prateek,
Question... why not in the OvsdbNodeCreateCommand where all of the other operational info is being created and written?
Ed
On Wed, Apr 22, 2015 at 3:22 PM, Prateek Garg (prategar)
<prategar@...> wrote:
Hi Ed,
The way I am implementing this is:
In OvsdbConnectionManager’s connected method, I’ll set system-Id from external_ids of open_Vswitch table in ConnectionInfo key (I’ll modify yang to add systemId field in connectionInfo to add system-id). Therefore to get the system-id in OvsdbConnectionManager’s
connected method, I need to execute select operation on open_Vswitch table.
So later when OvsdbNodeCreateCommand is called with ConnectionInfo, I’ll have the systemId to write to operation store.
Please let me know your thoughts on this approach.
On Wed, Apr 22, 2015 at 2:54 PM, Prateek Garg (prategar)
<prategar@...> wrote:
Hi,
While implementing this use case I need some help using the select operation of ovsdb library. I am trying to read the Open_vSwitch table to get the data in external-ids. I am doing the following:
Subject: Re: [ovsdb-dev] Arbitrary connection port is part of the node-id
The config/operational synch is best handled a little differently.
When you try to config an ovsdb-node to connect... you don't *really* know the systemid... just the IP:port.
The way we solved this for bridges, and the way I'd advocate solving it here is:
1) If you are configing a ovsdb-node, write its iid to the external-ids of the openvswitch table
2) When you get an openvswitch table, if it has an iid in its external-ids, use that in the operational data store as its iid
3) If you don't get an iid in its external-ids... use some reasonable heuristic to contruct one.
For #3... I don't currently see a reason *not* to use systemid presuming its guaranteed unique and we contruct some kind of reasonable
Uri from it.
Ed
On Tue, Apr 21, 2015 at 1:09 PM, Prateek Garg (prategar)
<prategar@...> wrote:
Hi,
In my opinion, the we can replace the node-id of the ovsdb node in md-sal to contain system_id from external properties. This would make sure that the config and operation store are in synch in case the controller reconnects
to the same ovs in tcp mode.
I’ll make these changes, update postman collection and submit for approval.
I am new to the OVSDB project. I need to get to speed as fast as I can from a developer view point (I mean eventually get into the code). So, what you steps you guys recommend for me to take?
Thanks much.
Khal
Re: OVSDB SB system test and scalability (if we can)
From: "Sharon Aicler (saichler)" <saichler@...> To: "Marcus G Koontz" <marcus.g.koontz@...>, "Jamo Luhrsen" <james.luhrsen@...>, "Luis Gomez" <ecelgp@...> Cc: "Praveen Mala" <praveen.mala@...>, "ovsdb-dev" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...>, "Sandeep Gangadharan (HP Networking)" <sandeep.gangadharan@...> Sent: Wednesday, May 13, 2015 7:05:31 PM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Hi All, I just performed some manual scale tests and found some hot spots i would like to share:
Thanks for doing this. This is good info to have.
I tried to connect an OVS with one bridge configured with 1K ports and found that it hits the limit of the OvsdbConnectonService instantiating the JsonRpcDecoder with a 100K size. If i change that to 1M it works so need to change this value to being configurable.
We have a known bug for this: https://bugs.opendaylight.org/show_bug.cgi?id=2732. It would be good to capture your details in there. - Do you have a practical deployment where you see something like this needed? - Did you testing show any particular piece of data creating large messages or was it just the fact that there were 1000 updates in a message? I know our Southbound adds the IID keys now and they are rather large.
Another hotspot is the notification scope from the ovsdb, because ovsdb southbound plugin requests all columns for all tables we receive notification with all the records every 2-3 seconds... which is a lot of data every 2-3 seconds to parse if you have 1K ports...or even 100 ports. I did some experiment and found that omitting the "statistics" & the "_version" columns from the register request in the method "monitorAllTables" class "OvsdbConnectionInstance" for all the tables is mitigating this issue and we get a notification only when an actual interesting change happens... Any reason why shouldn't we hard code to disregard those columns? I know statistics are collected via the OpenFlow StatisticsManager... (and _version is changed due to the change in statistics...)
Historically we monitor everything to ensure we don't miss anything - even stuff we care about and asked for explicitly. Sometimes ovsdb might change something and not send a notification that we expected. This seemed to change from version to version also. This is an item that has always been on the list to put in place though.
Could you file a bz and push gerrits for the changes you made so that we can track this effort and use your experiments?
Thanks, Sharon. ________________________________________ From: ovsdb-dev-bounces@... [ovsdb-dev-bounces@...] on behalf of Koontz, Marcus G [marcus.g.koontz@...] Sent: Tuesday, May 12, 2015 3:36 PM To: Luhrsen, Jamo; Luis Gomez Cc: Mala, Praveen; ovsdb-dev; Gangadharan, Sandeep (HP Networking); Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I'm free right now, I was on the OVSDB call. Or if tomorrow is better?
-----Original Message----- From: Luhrsen, Jamo [mailto:james.luhrsen@...] Sent: Tuesday, May 12, 2015 2:30 PM To: Luis Gomez Cc: Koontz, Marcus G; Neelima Sharma; Chaudhry usama; ovsdb-dev; Sam Hague; Mala, Praveen; Gangadharan, Sandeep (HP Networking) Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I’m good with a hangout until about 4:15pm PST today and tomorrow is wide open except for 10-11am PST.
JamO
On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
"Statistics" part i knew that it's going to stress southbound plugin, so we can for sure optimize it, until and unless someone comes up with the use case for using those statistics.
JsonRpcDecoder one is something i never hit, but we should fix this as well.
Created following cards for tracking these issues :
On Thu, May 14, 2015 at 4:35 AM, Sharon Aicler (saichler) <saichler@...> wrote:
Hi All,
I just performed some manual scale tests and found some hot spots i would like to share:
I tried to connect an OVS with one bridge configured with 1K ports and found that it hits the limit of the OvsdbConnectonService instantiating the JsonRpcDecoder with a 100K size. If i change that to 1M it works so need to change this value to being configurable.
Another hotspot is the notification scope from the ovsdb, because ovsdb southbound plugin requests all columns for all tables we receive notification with all the records every 2-3 seconds... which is a lot of data every 2-3 seconds to parse if you have 1K ports...or even 100 ports. I did some experiment and found that omitting the "statistics" & the "_version" columns from the register request in the method "monitorAllTables" class "OvsdbConnectionInstance" for all the tables is mitigating this issue and we get a notification only when an actual interesting change happens... Any reason why shouldn't we hard code to disregard those columns? I know statistics are collected via the OpenFlow StatisticsManager... (and _version is changed due to the change in statistics...)
Thanks,
Sharon.
________________________________________
From: ovsdb-dev-bounces@... [ovsdb-dev-bounces@...] on behalf of Koontz, Marcus G [marcus.g.koontz@...]
Sent: Tuesday, May 12, 2015 3:36 PM
To: Luhrsen, Jamo; Luis Gomez
Cc: Mala, Praveen; ovsdb-dev; Gangadharan, Sandeep (HP Networking); Neelima Sharma
Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I'm free right now, I was on the OVSDB call. Or if tomorrow is better?
-----Original Message-----
From: Luhrsen, Jamo [mailto:james.luhrsen@...]
Sent: Tuesday, May 12, 2015 2:30 PM
To: Luis Gomez
Cc: Koontz, Marcus G; Neelima Sharma; Chaudhry usama; ovsdb-dev; Sam Hague; Mala, Praveen; Gangadharan, Sandeep (HP Networking)
Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I’m good with a hangout until about 4:15pm PST today and tomorrow is wide open except for 10-11am PST.
JamO
> On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
>
> Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
>
>> On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
>>
>> Marcus,
>>
>> is your scale test (robot) pushed? If so, I can look it over and
>> help if possible. Sandeep, Luis and I have spent some cycles
>> recently working with openflow switch and host scale tests, so it’s all top of mind right now.
>>
>> JamO
>>
>>
>>> On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
>>>
>>> Hi All,
>>>
>>> I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
>>>
>>> I'm debugging the following basic scalability test (Robot):
>>> Setup env (odl with OVSDB SB)
>>> Add OVS switches in groups of 10
>>> Verify addition, check mem, cpu etc.
>>> Repeat
>>>
>>> I plan to expand the test once its working in its basic form. Some first expansion ideas:
>>> Cyclically add hosts as well
>>> Remove some switches before each new addition
>>> Verify some functionality between each addition
>>>
>>> I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
>>>
>>> I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
>>>
>>> Luis, Neelima and Chaudhry,
>>>
>>> I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium?
>>> I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
>>>
>>>
>>> Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
>>>
>>> Thanks,
>>> Marcus
>>>
>>> -----Original Message-----
>>> From: ovsdb-dev-bounces@...
>>> [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis
>>> Gomez
>>> Sent: Friday, May 8, 2015 2:48 PM
>>> To: Sam Hague; Chaudhry usama
>>> Cc: ovsdb-dev; Neelima Sharma
>>> Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we
>>> can)
>>>
>>> Adding Chaudhry (for some reason got removed)…
>>>
>>>
>>>> On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
>>>>
>>>> Hi Sam, couple of things:
>>>>
>>>> - Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
>>>>
>>>> - The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
>>>>
>>>> BR/Luis
>>>>
>>>>
>>>>
>>>>> On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
>>>>>
>>>>> Luis,
>>>>>
>>>>> Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
>>>>>
>>>>> It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
>>>>>
>>>>> Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
>>>>>
>>>>> Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
>>>>>
>>>>> Thanks, Sam
>>>>>
>>>>> ----- Original Message -----
>>>>>> From: "Anil Vishnoi" <vishnoianil@...>
>>>>>> To: "Luis Gomez" <ecelgp@...>
>>>>>> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma"
>>>>>> <neelima.sharma@...>
>>>>>> Sent: Friday, May 8, 2015 4:29:26 AM
>>>>>> Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if
>>>>>> we
>>>>>> can)
>>>>>>
>>>>>> Thanks Luis, this looks great!
>>>>>>
>>>>>>
>>>>>> Welcome!! Chaudhry, Neelima , please feel free to ping me
>>>>>> whenever you want to discussion any thing , my irc handle is
>>>>>> vishnoianil (or sometime i use avishnoi).
>>>>>>
>>>>>> Luis, About the docker testing, i think these instruction should
>>>>>> work, but i didn't try it personally. Sam/Flavio can probably comment on that.
>>>>>>
>>>>>> Thanks
>>>>>> Anil
>>>>>>
>>>>>> On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
>>>>>>
>>>>>>
>>>>>> Hi OVSDB devs,
>>>>>>
>>>>>> Please correct me if I am wrong but it is my understanding the
>>>>>> new OVSDB SB protocol is going to be consumed by several projects
>>>>>> in ODL (OpenStack, GBP, SFC, etc…). So the integration group
>>>>>> would like to help creating system test for the overlay API as
>>>>>> well as any other functionality you consider key in this release.
>>>>>> We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
>>>>>>
>>>>>> So if you agree on the above let me introduce you the 2 people
>>>>>> will be helping (starting next week):
>>>>>>
>>>>>> - Chaudhry Usama, OpenDaylight summer intern I mentor
>>>>>> - Neelima Sharma, freelance test engineer
>>>>>>
>>>>>> I am sure Chaudhry and Neelima will be very welcome in the OVSDB
>>>>>> community, one of the most open and diverse in OpenDaylight :)
>>>>>>
>>>>>> And now the basic questions to get started:
>>>>>>
>>>>>> 1) Information about new API, searching quickly:
>>>>>>
>>>>>> - I can see there is already this wiki under construction:
>>>>>> https://wiki.opendaylight.org/view/OVSDB:MDSAL_Southbound
>>>>>> - There is also a POSTMAN collection with supported methods:
>>>>>> https://github.com/opendaylight/ovsdb/blob/master/resources/commo
>>>>>> ns/ OVSDB_Southbound.postman_collection
>>>>>> - Some instruction on how to enable the SB feature and considerations:
>>>>>> https://wiki.opendaylight.org/view/OpenDaylight_OVSDB:Lithium_Int
>>>>>> egr
>>>>>> ation_Test
>>>>>> - Am I missing anything?
>>>>>>
>>>>>> 2) For scalability:
>>>>>>
>>>>>> - I remember old OVSDB committers were using docker instances to
>>>>>> encapsulate OVS and therefore saving lot of resources (multiple
>>>>>> OVS per host)
>>>>>> - I found this link:
>>>>>> https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do
>>>>>> you know if this is still updated?
>>>>>> - Any scalability/performance advice or concern is also welcome
>>>>>> :)
>>>>>>
>>>>>> Finally Chaudhry and Neelima have some experience with OVS but
>>>>>> they will need for sure some support with questions and issues when they start testing.
>>>>>> They both are in India so the time for the OVSDB call is not the
>>>>>> best but I will be in their behalf. I have also asked Anil to
>>>>>> support them in the time nobody (including me) is available.
>>>>>> Obviously I will be mentoring them in all that is test/automation
>>>>>> as well as with my limited OVS knowledge so I do not expect them
>>>>>> to ask questions all the time to OVSDB devs that I know they are very busy right now.
>>>>>>
>>>>>> And this is all I wanted to say, please let me know if you have
>>>>>> questions, suggestions, or anything.
>>>>>>
>>>>>> BR/Luis
>>>>>> _______________________________________________
>>>>>> ovsdb-dev mailing list
>>>>>> ovsdb-dev@...
>>>>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Thanks
>>>>>> Anil
>>>>>>
>>>>>> _______________________________________________
>>>>>> ovsdb-dev mailing list
>>>>>> ovsdb-dev@...
>>>>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>>>>
>>>>
>>>
>>> _______________________________________________
>>> ovsdb-dev mailing list
>>> ovsdb-dev@...
>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>
>
Hi Sharon, thanks for this effort. I am not a developer but it seems with the existing notification interval we may hit scalability issues, the same we had with openflowplugin stats in the beginning. Good news is that at least Anil and Ed know about this from openflow so they should be able to come up with a good suggestion.
On May 13, 2015, at 4:05 PM, Sharon Aicler (saichler) <saichler@...> wrote:
Hi All, I just performed some manual scale tests and found some hot spots i would like to share:
I tried to connect an OVS with one bridge configured with 1K ports and found that it hits the limit of the OvsdbConnectonService instantiating the JsonRpcDecoder with a 100K size. If i change that to 1M it works so need to change this value to being configurable.
Another hotspot is the notification scope from the ovsdb, because ovsdb southbound plugin requests all columns for all tables we receive notification with all the records every 2-3 seconds... which is a lot of data every 2-3 seconds to parse if you have 1K ports...or even 100 ports. I did some experiment and found that omitting the "statistics" & the "_version" columns from the register request in the method "monitorAllTables" class "OvsdbConnectionInstance" for all the tables is mitigating this issue and we get a notification only when an actual interesting change happens... Any reason why shouldn't we hard code to disregard those columns? I know statistics are collected via the OpenFlow StatisticsManager... (and _version is changed due to the change in statistics...)
Thanks, Sharon. ________________________________________ From: ovsdb-dev-bounces@... [ovsdb-dev-bounces@...] on behalf of Koontz, Marcus G [marcus.g.koontz@...] Sent: Tuesday, May 12, 2015 3:36 PM To: Luhrsen, Jamo; Luis Gomez Cc: Mala, Praveen; ovsdb-dev; Gangadharan, Sandeep (HP Networking); Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I'm free right now, I was on the OVSDB call. Or if tomorrow is better?
-----Original Message----- From: Luhrsen, Jamo [mailto:james.luhrsen@...] Sent: Tuesday, May 12, 2015 2:30 PM To: Luis Gomez Cc: Koontz, Marcus G; Neelima Sharma; Chaudhry usama; ovsdb-dev; Sam Hague; Mala, Praveen; Gangadharan, Sandeep (HP Networking) Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I’m good with a hangout until about 4:15pm PST today and tomorrow is wide open except for 10-11am PST.
JamO
On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
Hi All, I just performed some manual scale tests and found some hot spots i would like to share:
I tried to connect an OVS with one bridge configured with 1K ports and found that it hits the limit of the OvsdbConnectonService instantiating the JsonRpcDecoder with a 100K size. If i change that to 1M it works so need to change this value to being configurable.
Another hotspot is the notification scope from the ovsdb, because ovsdb southbound plugin requests all columns for all tables we receive notification with all the records every 2-3 seconds... which is a lot of data every 2-3 seconds to parse if you have 1K ports...or even 100 ports. I did some experiment and found that omitting the "statistics" & the "_version" columns from the register request in the method "monitorAllTables" class "OvsdbConnectionInstance" for all the tables is mitigating this issue and we get a notification only when an actual interesting change happens... Any reason why shouldn't we hard code to disregard those columns? I know statistics are collected via the OpenFlow StatisticsManager... (and _version is changed due to the change in statistics...)
Thanks, Sharon. ________________________________________ From: ovsdb-dev-bounces@... [ovsdb-dev-bounces@...] on behalf of Koontz, Marcus G [marcus.g.koontz@...] Sent: Tuesday, May 12, 2015 3:36 PM To: Luhrsen, Jamo; Luis Gomez Cc: Mala, Praveen; ovsdb-dev; Gangadharan, Sandeep (HP Networking); Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I'm free right now, I was on the OVSDB call. Or if tomorrow is better?
-----Original Message----- From: Luhrsen, Jamo [mailto:james.luhrsen@...] Sent: Tuesday, May 12, 2015 2:30 PM To: Luis Gomez Cc: Koontz, Marcus G; Neelima Sharma; Chaudhry usama; ovsdb-dev; Sam Hague; Mala, Praveen; Gangadharan, Sandeep (HP Networking) Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I’m good with a hangout until about 4:15pm PST today and tomorrow is wide open except for 10-11am PST.
JamO
On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
I would like to share with you the problems I encountered when I updated my PoC to ODL Helium. This PoC is currently built on: - OpenStack Juno (odl mechanism driver + dhcp agent + l3 agent), - ODL Helium (odl-obsdb-openstack and odl-ovsdb-northbound).
I hope this could help.
1. Race condition during the deleted interface notifications
ODL tries getting the neutron port (and next the neutron network) related to the deleted interface. This is a race condition because Neutron or Nova agents can send the delete orders before effectively deleting the interfaces from br-int. It results in remaining OpenFlow13 rules.
A possible bypass could be to search the associated local forward flow rules and get network types, ids and mac addresses. But that would oblige these keys to be simplified (only composed of the local port).
When neutron ports are being deleted, ODL deletes the interfaces from br-int. I think they must be managed by Openstack Neutron and Nova agents in charge of them.
Setting the OpenFlow13 rules required for my demo scenario takes now more than 1 minute instead of only few seconds before the upgrade. I think this results from the settings of Thread.sleep(500) in AbstractServiceInstace.java (everything works well again after decreasing this timer).
4. Additional data needed during interface updated notifications
In some special cases (e.g. detached ports), the external ids of an interface could be cleared. As processRowUpdate (see SouthboundHandler.java) forbids getting the old values, it also results in remaining OpenFlow13 rules.
5. Datapathid changes
This is related to my demo scenario when I tried bypassing the interface update processing issue. Here br-int's datapathid changes when eth1.100 is added (or deleted) to br-int (Linux 3.16, Open vSwitch 2.3.0+git20140819). As the datapathid changes, the OpenFlow13 rules cannot be deleted.
Best regards, Cédric
Re: OVSDB SB system test and scalability (if we can)
-----Original Message----- From: Luhrsen, Jamo [mailto:james.luhrsen@...] Sent: Tuesday, May 12, 2015 2:30 PM To: Luis Gomez Cc: Koontz, Marcus G; Neelima Sharma; Chaudhry usama; ovsdb-dev; Sam Hague; Mala, Praveen; Gangadharan, Sandeep (HP Networking) Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
I’m good with a hangout until about 4:15pm PST today and tomorrow is wide open except for 10-11am PST.
JamO
On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
On May 12, 2015, at 1:49 PM, Luis Gomez <ecelgp@...> wrote:
Crap, I just realized I missed the OVSDB call… Marcus, can we sync you, Jamo and I (and anybody interested) in a hangout?
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
On May 11, 2015, at 12:56 PM, Luhrsen, Jamo <james.luhrsen@...> wrote:
Marcus,
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
JamO
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
is your scale test (robot) pushed? If so, I can look it over and help if possible. Sandeep, Luis and I have spent some cycles recently working with openflow switch and host scale tests, so it’s all top of mind right now.
On May 11, 2015, at 11:04 AM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks, Marcus
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
On May 11, 2015, at 11:17 AM, Chaudhry usama <chaudhryusama@...> wrote:
Hi Marcus good to here a solid progress in scalability test. Don't have specific background so apologies.
Regarding tests which we are looking to start is to do as much functional testing of OVSDB southbound API that we can do till the lithium release. Yes we are shooting for lithium release.
On Mon, May 11, 2015 at 11:04 PM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot):
Setup env (odl with OVSDB SB)
Add OVS switches in groups of 10
Verify addition, check mem, cpu etc.
Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas:
Cyclically add hosts as well
Remove some switches before each new addition
Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium?
I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks,
Marcus
-----Original Message-----
From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez
Sent: Friday, May 8, 2015 2:48 PM
To: Sam Hague; Chaudhry usama
Cc: ovsdb-dev; Neelima Sharma
Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
> On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
>
> Hi Sam, couple of things:
>
> - Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
>
> - The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
>
> BR/Luis
>
>
>
>> On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
>>
>> Luis,
>>
>> Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
>>
>> It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
>>
>> Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
>>
>> Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
>>
>> Thanks, Sam
>>
>> ----- Original Message -----
>>> From: "Anil Vishnoi" <vishnoianil@...>
>>> To: "Luis Gomez" <ecelgp@...>
>>> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma"
>>> <neelima.sharma@...>
>>> Sent: Friday, May 8, 2015 4:29:26 AM
>>> Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we
>>> can)
>>>
>>> Thanks Luis, this looks great!
>>>
>>>
>>> Welcome!! Chaudhry, Neelima , please feel free to ping me whenever
>>> you want to discussion any thing , my irc handle is vishnoianil (or
>>> sometime i use avishnoi).
>>>
>>> Luis, About the docker testing, i think these instruction should
>>> work, but i didn't try it personally. Sam/Flavio can probably comment on that.
>>>
>>> Thanks
>>> Anil
>>>
>>> On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
>>>
>>>
>>> Hi OVSDB devs,
>>>
>>> Please correct me if I am wrong but it is my understanding the new
>>> OVSDB SB protocol is going to be consumed by several projects in ODL
>>> (OpenStack, GBP, SFC, etc…). So the integration group would like to
>>> help creating system test for the overlay API as well as any other
>>> functionality you consider key in this release. We would like to
>>> also include an scalability test for OVS instances and tunnel creation if time allows.
>>>
>>> So if you agree on the above let me introduce you the 2 people will
>>> be helping (starting next week):
>>>
>>> - Chaudhry Usama, OpenDaylight summer intern I mentor
>>> - Neelima Sharma, freelance test engineer
>>>
>>> I am sure Chaudhry and Neelima will be very welcome in the OVSDB
>>> community, one of the most open and diverse in OpenDaylight :)
>>>
>>> And now the basic questions to get started:
>>>
>>> 1) Information about new API, searching quickly:
>>>
>>> - I can see there is already this wiki under construction:
>>> https://wiki.opendaylight.org/view/OVSDB:MDSAL_Southbound
>>> - There is also a POSTMAN collection with supported methods:
>>> https://github.com/opendaylight/ovsdb/blob/master/resources/commons/
>>> OVSDB_Southbound.postman_collection
>>> - Some instruction on how to enable the SB feature and considerations:
>>> https://wiki.opendaylight.org/view/OpenDaylight_OVSDB:Lithium_Integr
>>> ation_Test
>>> - Am I missing anything?
>>>
>>> 2) For scalability:
>>>
>>> - I remember old OVSDB committers were using docker instances to
>>> encapsulate OVS and therefore saving lot of resources (multiple OVS
>>> per host)
>>> - I found this link:
>>> https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do
>>> you know if this is still updated?
>>> - Any scalability/performance advice or concern is also welcome :)
>>>
>>> Finally Chaudhry and Neelima have some experience with OVS but they
>>> will need for sure some support with questions and issues when they start testing.
>>> They both are in India so the time for the OVSDB call is not the
>>> best but I will be in their behalf. I have also asked Anil to
>>> support them in the time nobody (including me) is available.
>>> Obviously I will be mentoring them in all that is test/automation as
>>> well as with my limited OVS knowledge so I do not expect them to ask
>>> questions all the time to OVSDB devs that I know they are very busy right now.
>>>
>>> And this is all I wanted to say, please let me know if you have
>>> questions, suggestions, or anything.
>>>
>>> BR/Luis
>>> _______________________________________________
>>> ovsdb-dev mailing list
>>> ovsdb-dev@...
>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>
>>>
>>>
>>> --
>>> Thanks
>>> Anil
>>>
>>> _______________________________________________
>>> ovsdb-dev mailing list
>>> ovsdb-dev@...
>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>
>
Hi Marcus good to here a solid progress in scalability test. Don't have specific background so apologies.
Regarding tests which we are looking to start is to do as much functional testing of OVSDB southbound API that we can do till the lithium release. Yes we are shooting for lithium release.
On Mon, May 11, 2015 at 11:04 PM, Koontz, Marcus G <marcus.g.koontz@...> wrote:
Hi All,
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot):
Setup env (odl with OVSDB SB)
Add OVS switches in groups of 10
Verify addition, check mem, cpu etc.
Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas:
Cyclically add hosts as well
Remove some switches before each new addition
Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium?
I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
Thanks,
Marcus
-----Original Message-----
From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez
Sent: Friday, May 8, 2015 2:48 PM
To: Sam Hague; Chaudhry usama
Cc: ovsdb-dev; Neelima Sharma
Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
> On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
>
> Hi Sam, couple of things:
>
> - Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
>
> - The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
>
> BR/Luis
>
>
>
>> On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
>>
>> Luis,
>>
>> Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
>>
>> It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
>>
>> Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
>>
>> Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
>>
>> Thanks, Sam
>>
>> ----- Original Message -----
>>> From: "Anil Vishnoi" <vishnoianil@...>
>>> To: "Luis Gomez" <ecelgp@...>
>>> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma"
>>> <neelima.sharma@...>
>>> Sent: Friday, May 8, 2015 4:29:26 AM
>>> Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we
>>> can)
>>>
>>> Thanks Luis, this looks great!
>>>
>>>
>>> Welcome!! Chaudhry, Neelima , please feel free to ping me whenever
>>> you want to discussion any thing , my irc handle is vishnoianil (or
>>> sometime i use avishnoi).
>>>
>>> Luis, About the docker testing, i think these instruction should
>>> work, but i didn't try it personally. Sam/Flavio can probably comment on that.
>>>
>>> Thanks
>>> Anil
>>>
>>> On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
>>>
>>>
>>> Hi OVSDB devs,
>>>
>>> Please correct me if I am wrong but it is my understanding the new
>>> OVSDB SB protocol is going to be consumed by several projects in ODL
>>> (OpenStack, GBP, SFC, etc…). So the integration group would like to
>>> help creating system test for the overlay API as well as any other
>>> functionality you consider key in this release. We would like to
>>> also include an scalability test for OVS instances and tunnel creation if time allows.
>>>
>>> So if you agree on the above let me introduce you the 2 people will
>>> be helping (starting next week):
>>>
>>> - Chaudhry Usama, OpenDaylight summer intern I mentor
>>> - Neelima Sharma, freelance test engineer
>>>
>>> I am sure Chaudhry and Neelima will be very welcome in the OVSDB
>>> community, one of the most open and diverse in OpenDaylight :)
>>>
>>> And now the basic questions to get started:
>>>
>>> 1) Information about new API, searching quickly:
>>>
>>> - I can see there is already this wiki under construction:
>>> https://wiki.opendaylight.org/view/OVSDB:MDSAL_Southbound
>>> - There is also a POSTMAN collection with supported methods:
>>> https://github.com/opendaylight/ovsdb/blob/master/resources/commons/
>>> OVSDB_Southbound.postman_collection
>>> - Some instruction on how to enable the SB feature and considerations:
>>> https://wiki.opendaylight.org/view/OpenDaylight_OVSDB:Lithium_Integr
>>> ation_Test
>>> - Am I missing anything?
>>>
>>> 2) For scalability:
>>>
>>> - I remember old OVSDB committers were using docker instances to
>>> encapsulate OVS and therefore saving lot of resources (multiple OVS
>>> per host)
>>> - I found this link:
>>> https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do
>>> you know if this is still updated?
>>> - Any scalability/performance advice or concern is also welcome :)
>>>
>>> Finally Chaudhry and Neelima have some experience with OVS but they
>>> will need for sure some support with questions and issues when they start testing.
>>> They both are in India so the time for the OVSDB call is not the
>>> best but I will be in their behalf. I have also asked Anil to
>>> support them in the time nobody (including me) is available.
>>> Obviously I will be mentoring them in all that is test/automation as
>>> well as with my limited OVS knowledge so I do not expect them to ask
>>> questions all the time to OVSDB devs that I know they are very busy right now.
>>>
>>> And this is all I wanted to say, please let me know if you have
>>> questions, suggestions, or anything.
>>>
>>> BR/Luis
>>> _______________________________________________
>>> ovsdb-dev mailing list
>>> ovsdb-dev@...
>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>
>>>
>>>
>>> --
>>> Thanks
>>> Anil
>>>
>>> _______________________________________________
>>> ovsdb-dev mailing list
>>> ovsdb-dev@...
>>> https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev
>>>
>
I've been working on OVSDB SB Robot Scalability tests (not function) for a few weeks (since hackfest). My apologies for not updating my status. I've been pulled in a thousand directions and we've had a change of personnel committed to this effort. I was ramping another of my team to help in this space but they took a leave of absence. Now in addition to myself, Praveen Mala from my team at Intel will be helping out with this effort.
I'm debugging the following basic scalability test (Robot): Setup env (odl with OVSDB SB) Add OVS switches in groups of 10 Verify addition, check mem, cpu etc. Repeat
I plan to expand the test once its working in its basic form. Some first expansion ideas: Cyclically add hosts as well Remove some switches before each new addition Verify some functionality between each addition
I didn't see the docker info Luis mentioned until now, but had reached out to Dave Tucker awhile back with no response. I'll take a look and see if this can be incorporated in its current form. If the docker image needs to be updated I'd be willing to do that as well.
I've also been working on the Overlay code. I'd agree with Sam, Overlay is still in progress - system scalability and functionality tests are premature. The plan is to include basic tests once it is fully functional.
Luis, Neelima and Chaudhry,
I'd be glad to talk more so as not to duplicate effort. What kinds of tests are you looking at to start? What is the timeframe target for your work? Are you shooting for lithium? I'd love to get my scalability test in for Lithium but have been hampered by the larger than expected ramp involved and debug issues.
Sorry again for the radio silence on this, things have been a bit crazy for me the last few weeks.
-----Original Message----- From: ovsdb-dev-bounces@... [mailto:ovsdb-dev-bounces@...] On Behalf Of Luis Gomez Sent: Friday, May 8, 2015 2:48 PM To: Sam Hague; Chaudhry usama Cc: ovsdb-dev; Neelima Sharma Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Adding Chaudhry (for some reason got removed)…
On May 8, 2015, at 10:54 AM, Luis Gomez <ecelgp@...> wrote:
Hi Sam, couple of things:
- Can you refer any documentation or POSTMAN collection for the small OVSDB pieces you mention or there is really nothing (other than code) as you say?
- The OpenStack is a different track Daniel is driving and trying to align with OPNFV goals. That one I am not 100% certain we can get it done this release according to latest Daniel comments in the integration call. However if with your experience with tempest test you think we can easily deploy stable openstack with unstable ODL and make some quick VM transport test, we can also take a look at that now.
BR/Luis
On May 8, 2015, at 10:39 AM, Sam Hague <shague@...> wrote:
Luis,
Welcome Chaudhry and Neelima! Yeah, this is great and something that would be really helpful to have. You are in good hands with Anil.
It would be good to include the whole southbound API and not just the tunnel overlay. The tunnel overlay API is still a work in progress. The underlying tasks to create an overlay are implemented so many users are currently using that in the meantime. I think that is actually easier to test since it broken down into smaller, understandable pieces - add bridge, add tunnel port, add flows to make tunnel. The links you mentioned below are the only ones we have right now. It is probably good to look at the ovsdb.yang or the restconf explorer to see the model. We can elaborate on the different steps needed to build a tunnel.
Scalability is also another test that is greatly needed. I don't think we have been updating the docker so not sure if it still works but it shouldn't take much to get it there if it is broken. I would suggest giving it a shot to see.
Question: how does this relate to other work for OpenStack integration? I recall Daniel was working on something related and Patrick was trying to get some multi-node testing setup. I think Daniel also might have done some docker work that we could leverage.
Thanks, Sam
----- Original Message -----
From: "Anil Vishnoi" <vishnoianil@...> To: "Luis Gomez" <ecelgp@...> Cc: "<ovsdb-dev@...>" <ovsdb-dev@...>, "Neelima Sharma" <neelima.sharma@...> Sent: Friday, May 8, 2015 4:29:26 AM Subject: Re: [ovsdb-dev] OVSDB SB system test and scalability (if we can)
Thanks Luis, this looks great!
Welcome!! Chaudhry, Neelima , please feel free to ping me whenever you want to discussion any thing , my irc handle is vishnoianil (or sometime i use avishnoi).
Luis, About the docker testing, i think these instruction should work, but i didn't try it personally. Sam/Flavio can probably comment on that.
Thanks Anil
On Fri, May 8, 2015 at 8:26 AM, Luis Gomez < ecelgp@... > wrote:
Hi OVSDB devs,
Please correct me if I am wrong but it is my understanding the new OVSDB SB protocol is going to be consumed by several projects in ODL (OpenStack, GBP, SFC, etc…). So the integration group would like to help creating system test for the overlay API as well as any other functionality you consider key in this release. We would like to also include an scalability test for OVS instances and tunnel creation if time allows.
So if you agree on the above let me introduce you the 2 people will be helping (starting next week):
- Chaudhry Usama, OpenDaylight summer intern I mentor - Neelima Sharma, freelance test engineer
I am sure Chaudhry and Neelima will be very welcome in the OVSDB community, one of the most open and diverse in OpenDaylight :)
- I remember old OVSDB committers were using docker instances to encapsulate OVS and therefore saving lot of resources (multiple OVS per host) - I found this link: https://wiki.opendaylight.org/view/OVSDB:Testing_with_Docker , do you know if this is still updated? - Any scalability/performance advice or concern is also welcome :)
Finally Chaudhry and Neelima have some experience with OVS but they will need for sure some support with questions and issues when they start testing. They both are in India so the time for the OVSDB call is not the best but I will be in their behalf. I have also asked Anil to support them in the time nobody (including me) is available. Obviously I will be mentoring them in all that is test/automation as well as with my limited OVS knowledge so I do not expect them to ask questions all the time to OVSDB devs that I know they are very busy right now.
And this is all I wanted to say, please let me know if you have questions, suggestions, or anything.
I believe this is expected behaviour ( given that we don't do node reconciliation after controller restart). Step 3 below connect controller to ovs, but because clustering is enable, your config data persist across controller restart, and when you fire
the same request again ( step 3 ) after controller reboot, data store does not generate any data change event for that request because the exact same config already exist in config data store. And because southbound plugin don't get any data change notification
it does not initiate connection to ovs and thats the reason you don't see operational data in data store.
To make it work, you need to first make sure that you delete the existing configuration and add it again. I know this is not a good solution. I know two solution that i can think of 1) do node reconciliation at the controller restart by reading the existing
configuration 2) if clustered data store can notify about the existing configuration after controller reboot, and the ovsdb module can do node reconciliation based on that. Both the approach has their own pros/cons but i believe we need to live with t, till
we make the application fully cluster aware.
Hope it helps
Anil
On 10-May-2015, at 12:12 pm, Prateek Garg (prategar) <prategar@...> wrote:
I believe this is expected behaviour ( given that we don't do node reconciliation after controller restart). Step 3 below connect controller to ovs, but because clustering is enable, your config data persist across controller restart, and when you fire the same request again ( step 3 ) after controller reboot, data store does not generate any data change event for that request because the exact same config already exist in config data store. And because southbound plugin don't get any data change notification it does not initiate connection to ovs and thats the reason you don't see operational data in data store.
To make it work, you need to first make sure that you delete the existing configuration and add it again. I know this is not a good solution. I know two solution that i can think of 1) do node reconciliation at the controller restart by reading the existing configuration 2) if clustered data store can notify about the existing configuration after controller reboot, and the ovsdb module can do node reconciliation based on that. Both the approach has their own pros/cons but i believe we need to live with t, till we make the application fully cluster aware. Hope it helps
