Date
1 - 13 of 13
[OpenDaylight TSC] [OpenDaylight][TSC][release] Fluorine SR3 status 13/06 - RC2 candidate sign off
|
Robert Varga
On 15/06/2019 00:14, Luis Gomez wrote:
As I mentioned earlier, there has been some change in karaf fluorine soThis is probably related to BC 1.61, quoting 1.62 release notes: A performance issue with RSA key pair generation that was introduced in 1.61 has been mostly eliminated. I am not sure it's worth re-spinning the whole thing, though... Bye, R. |
|
|
|
JamO Luhrsen
On 6/15/19 4:16 AM, Robert Varga wrote:
On 15/06/2019 00:14, Luis Gomez wrote:is this a cumbersome process for us? Like we've decided before with FluorineAs I mentioned earlier, there has been some change in karaf fluorine soThis is probably related to BC 1.61, quoting 1.62 release notes: SR3, since it's the last release for Fluorine maybe we take the time to make sure we get it as polished as possible. The CSIT jobs to vet afterward are a smaller set than in previous releases so it's really only hopefully another 2-3 days to get it done once we have the BC upgrade. JamO
|
|
|
|
Robert Varga
On 17/06/2019 18:32, Jamo Luhrsen wrote:
AFAICT, this requires:This is probably related to BC 1.61, quoting 1.62 release notes:is this a cumbersome process for us? Like we've decided before with - releasing odlparent-3.1.8 - creating version bump patches for all projects - releasing yangtools-2.0.21 - creating version bump patches for all projects - merging both sets of patches - spinning autorelease The CSIT jobs to vet afterward are a smaller set than in previousRight, but I don't have the cycles to drive this, so unless someone else steps in... Regards, Robert |
|
|
|
Robert Varga
On 18/06/2019 14:47, Robert Varga wrote:
Right, but I don't have the cycles to drive this, so unless someone elseHonestly, given TSC's approval: AGREED: The TSC approves build 143 as the Fluorine SR3 build, pending CSIT jobs (abhijitk, 16:26:08) I would propose to just release SR3 and make it the TSC's decision on whether to release an SR4 with just this fix. Resource cost is *almost* the same -- excluding the version bump cost. Regards, Robert |
|
|
|
JamO Luhrsen
On 6/18/19 1:31 PM, Robert Varga wrote:
On 18/06/2019 14:47, Robert Varga wrote:That's fair and reasonable to me. We can discuss the SR4 option on ourRight, but I don't have the cycles to drive this, so unless someone elseHonestly, given TSC's approval: TSC meeting this week, and go ahead and get SR3 out the door. JamO
|
|
|
|
I'm not sure that we will have enough quorum on tomorrow's TSC meeting but we could definitely release SR3 as it is since customers are waiting.. Do we want to just move forward via email? On Tue, Jun 18, 2019 at 1:33 PM Jamo Luhrsen <jluhrsen@...> wrote:
|
|
|
|
Luis Gomez
I am fine to release as it is, we can maybe add a comment in the release notes mentioning the BC 1.61 perf issue and the impact in karaf console SSH.
toggle quoted message
Show quoted text
BR/Luis
|
|
|
|
JamO Luhrsen
In today's TSC call, Robert suggested a test where we just
replace two BC
toggle quoted message
Show quoted text
jars in Fluorine and test what the BC upgrade to 1.62 would look like. I can confirm the upgraded jars improve the initial key pair generation to under 10s which is the timeout our robot automation uses. specifically, I measured 6-8 seconds over several iterations of the test. Using 1.61, I measured 15-25 seconds over several iterations of the test. I think it's safe to say that a BC upgrade would fix our robot CSIT failures in Fluorine. Is it worth our time and effort to upgrade Fluorine and release an SR4? I briefly checked neon csit and did not see any case where we are hitting this problem there. We are still using BC1.61 in neon, but there was a an upgrade to our SSH in general after fluorine so it's likely that is working better with 1.61. Thanks, JamO On 6/20/19 12:31 AM, Luis Gomez wrote:
I am fine to release as it is, we can maybe add a comment in the release notes mentioning the BC 1.61 perf issue and the impact in karaf console SSH. |
|
|
|
Robert Varga
On 20/06/2019 23:12, Jamo Luhrsen wrote:
In today's TSC call, Robert suggested a test where we just replace two BCThanks for confirming. The request to release odlparent-3.1.8 with this upgrade is here: https://jira.linuxfoundation.org/servicedesk/customer/portal/2/IT-16534 Regards, Robert |
|
|
|
Robert Varga
On 21/06/2019 08:16, Robert Varga wrote:
Thanks for confirming. The request to release odlparent-3.1.8 with thisBoth odlparent-3.1.8 and yangtools-2.0.21 are out there, now it's up to someone to raise/merge the patches to integrate them -- just like https://git.opendaylight.org/gerrit/#/q/topic:mri-fluorine-sr3 did. Regards, Robert |
|
|
|
JamO Luhrsen
On 6/21/19 7:48 AM, Robert Varga wrote:
On 21/06/2019 08:16, Robert Varga wrote:working on it here:Thanks for confirming. The request to release odlparent-3.1.8 with thisBoth odlparent-3.1.8 and yangtools-2.0.21 are out there, now it's up to https://git.opendaylight.org/gerrit/#/q/topic:mri-fluorine-sr4 not done yet, but once I am what is the next step? I think it's the multipatch job right? but I've not worked through these things before so if anyone has some steps/docs for me, I'll follow them. JamO
|
|
|
|
Luis. See the email below from Jamo regarding flourine sr4. I don’t think it is worth it at this time but we can review it in detail on the next tsc On Thu, Jun 20, 2019 at 2:12 PM Jamo Luhrsen <jluhrsen@...> wrote:
--
|
|
|
|
Robert Varga
On 22/06/2019 02:23, Jamo Luhrsen wrote:
Not sure there is an explicit document. The next step is to run a multipatch -- running here: https://jenkins.opendaylight.org/releng/job/integration-multipatch-test-fluorine/298/ The job output needs to be check if it references any old versions and whether it passes... Regards, Robert |
|
|