On Fri, Sep 8, 2017 at 8:28 AM, Jozef Bacigál <jozef.baci...@pantheon.tech> wrote:
> Thanks Michael, thanks for information. :) > > > > I do investigations further for the Illegal state bug. > Jozef, Abhijit, Anil: Have you seen https://bugs.opendaylight.org/show_bug.cgi?id=9038#c19 - great news - thanks a lot, Sridhar and everyone involved! So... will you merge https://git.opendaylight.org/gerrit/#/c/62674/6 ASAP, cherry-pick it from stable/carbon to stable/nitrogen and master, and... done! > But at least we got some issue solved :) > > > > Jozef > > > > P.S.: Let you know if something new > > > > *From:* Michael Vorburger [mailto:vorbur...@redhat.com] > *Sent:* Thursday, September 7, 2017 3:07 PM > *To:* Jozef Bacigál <jozef.baci...@pantheon.tech>; Anil Vishnoi < > vishnoia...@gmail.com>; Vishnoi, Anil <avish...@brocade.com> > *Cc:* Robert Varga <n...@hq.sk>; Tomáš Slušný <tomas.slu...@pantheon.tech>; > Abhijit Kumbhare <abhijitk...@gmail.com>; openflowplugin-dev@lists. > opendaylight.org; controller-dev@lists.opendaylight.org; > mdsal-...@lists.opendaylight.org > *Subject:* Re: [mdsal-dev] [openflowplugin-dev] Fwd: Bug 9038 - > IllegalStateException: Attempted to close chain with outstanding > transaction PingPongTransaction at org.opendaylight. > openflowplugin.impl.device.TransactionChainManager.createTxChain > > > > All, > > > > On Wed, Sep 6, 2017 at 7:26 PM, Michael Vorburger <vorbur...@redhat.com> > wrote: > > On Wed, Sep 6, 2017 at 4:34 PM, Michael Vorburger <vorbur...@redhat.com> > wrote: > > On Wed, Sep 6, 2017 at 3:59 PM, Jozef Bacigál <jozef.baci...@pantheon.tech> > wrote: > > HI Michael, > > are you able to test it on this patch [1] ? Or it needed to be merged ? > > Jozef - thanks a lot! I'll discuss with Daniel to understand how easy/hard > it would be to get a carbon RPM with this, and let you know ASAP. > > > > Jozef, FYI we've just been able to get a new full ODL dist build + RPM > incl. this (unmerge) change, and will re-run this through our scale test > environment ASAP, and let you know how it looks. > > > > There is Good and Bad news - which one do you want first? ;-) > > The Good: I’m (very) happy to hereby confirm that on a re-test of a custom > distribution of latest (yesterday's) carbon/stable together with c/62674/5 > (NB Patch Set 5; it’s still WIP) pulled in, we do not see > https://bugs.opendaylight.org/show_bug.cgi?id=9034 (OOM due to huge Map > in ShardDataTree) anymore! This is great. NB: It may have been c/62674 or > anything else that went in recently (anywhere incl. Bug 9028, dunno if that > could have helped for this?) which did the trick, but it works now, so > great! FYI this has been verified by both "wow, now it didn't blew up > anymore with OOM" as well as confirmed via HPROF analysis through MAT > (which now also revealed new Bug 9114; but that's another story). I'll > therefore close blocker Bug 9034 as soon as you've agreed a final version > of c/62674 and merged it (into carbon, nitrogen and master). > > The FYI: c/62674/5 does fix https://bugs.opendaylight.org/ > show_bug.cgi?id=9070, so I'll also close that after you merge. > > > > The Bad: The https://bugs.opendaylight.org/show_bug.cgi?id=9038 > "IllegalStateException: Attempted to close chain with outstanding > transaction PingPongTransaction" still appears in logs - sorry. So c/62674/5 > *DOES NOT* fix blocker Bug 9038. You can also find related information on > the new attachment on umbrella https://bugs.opendaylight.org/ > show_bug.cgi?id=9096 re. other Tx leaks. I'm hoping this information is > useful to openflowplug-dev to find another solution for Bug 9038. > > > Tx, > > M. > > PS: Wrote about some blurb about the new tooling which came out of this on > http://blog2.vorburger.ch/2017/09/how-to-find-transaction-related-memory. > html > > -- > > Michael Vorburger, Red Hat > vorbur...@redhat.com | IRC: vorburger @freenode | ~ = http://vorburger.ch >
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev
