Hi Abhjit,

This is something I have denounced many times and I will be doing until it gets fixed: regardless the OpenFlow functionality, the amount of OSGi warnings and tracebacks we get when running the new ofplugin (master or stable branch) is overwhelming, so much that I agree with Madhu it is very difficult to identify real issues among so many logs. I understand this can be convenient for CI process and ofplugin developers but I think it is totally unacceptable for a service or any other ODL official release.

BR/Luis

On Jul 12, 2014, at 10:42 AM, Abhijit Kumbhare <abhijitkoss@...> wrote:

So Madhu (Brent in Madhu's absence) - about your comment:

===========================================

Starting the Virtualization edition using : ./run.sh -Xmx1024m -XX:MaxPermSize=384m -of13 -virt ovsdb

and immediately after connecting OF13 switches, I see a bunch of these Hydrogen trademark tracebacks :
https://gist.github.com/fdb32280f3079cf88d77

I thought the Stable release should have taken care of these tracebacks ?
Without any proper error message, it is tough to find out what causes these.

I believe it is important to get these resolved in the stable/hydrogen branch. 

==============================================

Are these tracebacks not present in the master? It may be tough to track down the change in the master in the next 2-3 days & get it tested - in time for the next week's release. The questions are is it affecting the functionality? Is it a regression from Hydrogen? Depending on the answers - is this something that can be fixed in a next Hydrogen stable release? 

Added all 3 projects - controller, openflow-java and openflow-plugin if they have more insight on the errors.

On Sat, Jul 12, 2014 at 3:30 AM, Madhu Venugopal <mavenugo@...> wrote:

All,

Used the following image :

https://jenkins.opendaylight.org/odlautorelease/job/autorelease-major-stablehydrogen/39/artifact/autorelease/stagingreleases/m2repo/org/opendaylight/integration/distributions-virtualization/0.1.1-1/

Starting the Virtualization edition using : ./run.sh -Xmx1024m -XX:MaxPermSize=384m -of13 -virt ovsdb

and immediately after connecting OF13 switches, I see a bunch of these Hydrogen trademark tracebacks :
https://gist.github.com/fdb32280f3079cf88d77

I thought the Stable release should have taken care of these tracebacks ?
Without any proper error message, it is tough to find out what causes these.
I believe it is important to get these resolved in the stable/hydrogen branch.

But the basic L2 ping in ovsdb + of13 based Openstack - vxlan overlay setup is working.

Unfortunately, the LLDP issue that we saw in Hydrogen release is still seen here.
The fix : https://git.opendaylight.org/gerrit/#/c/7144/ in ovsdb project is supposed to have been ported over to stable/hydrogen
but slipped through the cracks. But, This is NOT a release-blocker.

I will be MIA starting today for 1 week and hence have copied Brent & ovsdb-dev team to port this bug over to the stable/hydrogen
branch if in case, there is another respin is in the works to address the above tracebacks issue.

Thanks,
Madhu

On 7/11/14, 5:21 PM, Ryan Moats wrote:

All-

Because this only went out today, the deadline dates from #1 and 2 below are changing by about 30 hours:

1) Folks have until 2330 UTC/1630 PDT 7/12 to respond to this mail with pointers to merged patches that should be included in the next stable/Hydrogren RC.
Note: we have already started the RC production process, so if there is no response to this email, we will use the output of that process...

2) Once the RC is cut, email will be sent and folks have until 2330 UTC/1630 PDT 7/16 to test and make an objection if another RC is needed.

3) The 7/17 TOC date doesn't change.

Ryan Moats (regXboi)

Chris Wright <chrisw@...> wrote on 07/11/2014 06:24:44 PM:

> From: Chris Wright <chrisw@...>
> To: "Dana Kutenicsova -X (dkutenic - Pantheon Technologies SRO at
> Cisco)" <dkutenic@...>, David Goldberg
> <David.Goldberg@...>, Luis Gomez <ecelgp@...>,
> Ryan Moats/Omaha/IBM@IBMUS, Abhijit Kumbhare
> <abhijitkoss@...>, Chris Wright <chrisw@...>, "Leena
> Ratnam (lratnam)" <lratnam@...>, "derick.winkworth@..."
> <derick.winkworth@...>, "Ed Warnicke (eaw)" <eaw@...>,
> "gerag@..." <gerag@...>, Hideyuki Tai <h-
> tai@...>, Madhu Venugopal <mavenugo@...>, Michal
> Polkoráb <michal.polkorab@...>, "Sarath Babu G
> (sarathbg@...)" <sarathbg@...>, "Tony Tkacik -X (ttkacik -
> Pantheon Technologies SRO at Cisco)" <ttkacik@...>, Anil
> Vishnoi <vishnoianil@...>, �x宜玲 <ylhsieh@...>, Gal
> Mainzer <GMainzer@...>
> Cc: release@..., tsc@...
> Date: 07/11/2014 06:25 PM
> Subject: Hydrogen Stable release candidate update
>
> (Ugh, I just found I never sent this!)
>
> Hi All,
>
> In response to Dana's email re: BGPCEP and the resulting discussing on
> today's TSC meeting, we've agreed to this plan:
>
> 1) We will cut updated artifacts tomorrow, roughly 24 hours from now,
> July 11th at 18:00 UTC/11am PDT.
>
> 2) Dana's note identified changes needed in yangtools, so we expect
> all projects should run their full regression testing against the new
> artifacts.  Please report your results no later than Tue, July 15th at
> 18:00 UTC/11am PDT.
>
> If you discover new regressions (notably, those caused by the recent
> changes to yangtools) we will need to consider the issues case by case.
>
> If all projects are passing cleanly then we can consider this the final
> release candidate and...
>
> 3) Next week's TSC meeting, July 17th at 17:00UTC/10am PDT, we will
> decide whether to release those artifacts.
>
> thanks,
> -chris
>

_______________________________________________
controller-dev mailing list
controller-dev@...
https://lists.opendaylight.org/mailman/listinfo/controller-dev

Thanks Abhijit, 

Since we need to respin the stable release due to last-minute patches in BGP/PCEP project, do you think it is feasible to do something quick about below issues?

BR/Luis

 

On Jul 14, 2014, at 1:18 PM, Abhijit Kumbhare <abhijitkoss@...> wrote:

Luis,

I agree with you.

Abhijit

On Sat, Jul 12, 2014 at 2:11 PM, Luis Gomez <ecelgp@...> wrote:

Hi Abhjit,

This is something I have denounced many times and I will be doing until it gets fixed: regardless the OpenFlow functionality, the amount of OSGi warnings and tracebacks we get when running the new ofplugin (master or stable branch) is overwhelming, so much that I agree with Madhu it is very difficult to identify real issues among so many logs. I understand this can be convenient for CI process and ofplugin developers but I think it is totally unacceptable for a service or any other ODL official release.

BR/Luis

On Jul 12, 2014, at 10:42 AM, Abhijit Kumbhare <abhijitkoss@...> wrote:

So Madhu (Brent in Madhu's absence) - about your comment:

===========================================

Starting the Virtualization edition using : ./run.sh -Xmx1024m -XX:MaxPermSize=384m -of13 -virt ovsdb

and immediately after connecting OF13 switches, I see a bunch of these Hydrogen trademark tracebacks :
https://gist.github.com/fdb32280f3079cf88d77

I thought the Stable release should have taken care of these tracebacks ?
Without any proper error message, it is tough to find out what causes these.

I believe it is important to get these resolved in the stable/hydrogen branch. 

==============================================

Are these tracebacks not present in the master? It may be tough to track down the change in the master in the next 2-3 days & get it tested - in time for the next week's release. The questions are is it affecting the functionality? Is it a regression from Hydrogen? Depending on the answers - is this something that can be fixed in a next Hydrogen stable release? 

Added all 3 projects - controller, openflow-java and openflow-plugin if they have more insight on the errors.

On Sat, Jul 12, 2014 at 3:30 AM, Madhu Venugopal <mavenugo@...> wrote:

All,

Used the following image :

https://jenkins.opendaylight.org/odlautorelease/job/autorelease-major-stablehydrogen/39/artifact/autorelease/stagingreleases/m2repo/org/opendaylight/integration/distributions-virtualization/0.1.1-1/

Starting the Virtualization edition using : ./run.sh -Xmx1024m -XX:MaxPermSize=384m -of13 -virt ovsdb

and immediately after connecting OF13 switches, I see a bunch of these Hydrogen trademark tracebacks :
https://gist.github.com/fdb32280f3079cf88d77

I thought the Stable release should have taken care of these tracebacks ?
Without any proper error message, it is tough to find out what causes these.
I believe it is important to get these resolved in the stable/hydrogen branch.

But the basic L2 ping in ovsdb + of13 based Openstack - vxlan overlay setup is working.

Unfortunately, the LLDP issue that we saw in Hydrogen release is still seen here.
The fix : https://git.opendaylight.org/gerrit/#/c/7144/ in ovsdb project is supposed to have been ported over to stable/hydrogen
but slipped through the cracks. But, This is NOT a release-blocker.

I will be MIA starting today for 1 week and hence have copied Brent & ovsdb-dev team to port this bug over to the stable/hydrogen
branch if in case, there is another respin is in the works to address the above tracebacks issue.

Thanks,
Madhu

On 7/11/14, 5:21 PM, Ryan Moats wrote:

All-

Because this only went out today, the deadline dates from #1 and 2 below are changing by about 30 hours:

1) Folks have until 2330 UTC/1630 PDT 7/12 to respond to this mail with pointers to merged patches that should be included in the next stable/Hydrogren RC.
Note: we have already started the RC production process, so if there is no response to this email, we will use the output of that process...

2) Once the RC is cut, email will be sent and folks have until 2330 UTC/1630 PDT 7/16 to test and make an objection if another RC is needed.

3) The 7/17 TOC date doesn't change.

Ryan Moats (regXboi)

Chris Wright <chrisw@...> wrote on 07/11/2014 06:24:44 PM:

> From: Chris Wright <chrisw@...>
> To: "Dana Kutenicsova -X (dkutenic - Pantheon Technologies SRO at
> Cisco)" <dkutenic@...>, David Goldberg
> <David.Goldberg@...>, Luis Gomez <ecelgp@...>,
> Ryan Moats/Omaha/IBM@IBMUS, Abhijit Kumbhare
> <abhijitkoss@...>, Chris Wright <chrisw@...>, "Leena
> Ratnam (lratnam)" <lratnam@...>, "derick.winkworth@..."
> <derick.winkworth@...>, "Ed Warnicke (eaw)" <eaw@...>,
> "gerag@..." <gerag@...>, Hideyuki Tai <h-
> tai@...>, Madhu Venugopal <mavenugo@...>, Michal
> Polkoráb <michal.polkorab@...>, "Sarath Babu G
> (sarathbg@...)" <sarathbg@...>, "Tony Tkacik -X (ttkacik -
> Pantheon Technologies SRO at Cisco)" <ttkacik@...>, Anil
> Vishnoi <vishnoianil@...>, �x宜玲 <ylhsieh@...>, Gal
> Mainzer <GMainzer@...>
> Cc: release@..., tsc@...
> Date: 07/11/2014 06:25 PM
> Subject: Hydrogen Stable release candidate update
>
> (Ugh, I just found I never sent this!)
>
> Hi All,
>
> In response to Dana's email re: BGPCEP and the resulting discussing on
> today's TSC meeting, we've agreed to this plan:
>
> 1) We will cut updated artifacts tomorrow, roughly 24 hours from now,
> July 11th at 18:00 UTC/11am PDT.
>
> 2) Dana's note identified changes needed in yangtools, so we expect
> all projects should run their full regression testing against the new
> artifacts.  Please report your results no later than Tue, July 15th at
> 18:00 UTC/11am PDT.
>
> If you discover new regressions (notably, those caused by the recent
> changes to yangtools) we will need to consider the issues case by case.
>
> If all projects are passing cleanly then we can consider this the final
> release candidate and...
>
> 3) Next week's TSC meeting, July 17th at 17:00UTC/10am PDT, we will
> decide whether to release those artifacts.
>
> thanks,
> -chris
>

_______________________________________________
controller-dev mailing list
controller-dev@...
https://lists.opendaylight.org/mailman/listinfo/controller-dev