Tuesday, 2026-01-13

mkurohaHi gibi, sorry to bother you just after you got back from PTO. I’ve updated this patch based on your feedback. Please take a look and share any comments when you have time. https://review.opendev.org/c/openstack/nova/+/94830402:39
*** mhen_ is now known as mhen02:55
seyeongkimcould someone please review this patch or give me any advices about the approch? thanks so much. https://review.opendev.org/c/openstack/nova/+/96366503:01
*** EugenMayer4401802 is now known as EugenMayer44018007:16
gibisambork: o/ I'm not sure you saw gmaan's post merge comments on the conductor eventlet series. https://review.opendev.org/c/openstack/nova/+/962478 I think it worth checking if there are dead code / missing cleanup based on his comments in that patch08:52
samborkhey gibi, thanks for pointing out this one! yeah I will check them !08:55
bauzasUggla: fancy mentioning today as implementation review day ? :)09:30
Ugglabauzas, you are right, I forgot about it yesterday.09:39
bauzasUggla: can I use https://etherpad.opendev.org/p/nova-2026.1-status as a base for reviewing patches09:40
bauzas?09:40
Ugglabauzas, I still not updated it (gonna do it). So unfortunately not. But you can have a look at r_taken's https://blueprints.launchpad.net/nova/+spec/generalize-sev-code patch09:43
bauzasthanks09:43
bauzasand I also said to tkajinam that I was wanting to review his patches, tkajinam any first ones I should look at ?09:43
bauzasUggla: can you please set https://blueprints.launchpad.net/nova/+spec/generalize-sev-code to approved as we agreed at the PTG ?09:47
Ugglabauzas, yep09:54
tkajinambauzas, all of my patches are in single chain so please start from https://review.opendev.org/c/openstack/nova/+/969263 (I triggered recheck because the failure in threading job looks unrelated (and has never seen before)11:04
tkajinamthe first one is the preparation work which extends the existing functional test coverage so I believe it's not very controversial  :-)11:05
sean-k-mooneyseyeongkim: so that is a feature rather then a bug. its alos possibel to specify the boot order via teh block device mappings today13:45
sean-k-mooneylookign at the bug report this does not seam to be a nova bug13:50
sean-k-mooneythe behvario depend on ght grub config and grub config is out os scope of nova/openstack13:50
sean-k-mooneyin nova we also dont offically suprpot dual booting.13:51
sean-k-mooneyat lest not in multipel disks like this13:51
sean-k-mooneywe supprot a singel root disk(local or cidner volume)  + additiaonl ephermal disks or cinder data volumes13:52
sean-k-mooneythe bug report state that if both device hav ea boot_index set then it works today13:52
sean-k-mooneyif you do not use cinder voluems for the non root disk that woudl also be lost on operations like shelve13:54
gibisean-k-mooney: fyi the py312-treading job error tkajinam mentioned above in https://review.opendev.org/c/openstack/nova/+/969263 is complaining to FairLockGuard. So you might have imediate context on that issue 16:25
gibiTypeError: Cannot enter FairLockGuard while it is already active. Create a new instance for nested usage or wait for the current context to exit.16:25
gibihttps://zuul.opendev.org/t/openstack/build/ee184acdd25a404d872dbd294af556d216:25
gmaangibi: sambork: more than dead code,  my main concern is cleaning up the scatter-gather executer. we can discuss it in wed call.16:44
gibigmaan: OK. I haven't read your comments deep enough. Indeed it is better to discuss on the tomorrows call.16:49
gmaan++16:49
sean-k-mooneygibi: i can take a look. its possibel that we are leakign locks between tests althoguh im not entirly sure how that woudl happen in this case.17:05
sean-k-mooneynova.tests.unit.test_utils.TestFairLockGuard.test_context_manager_shared_between_threads17:07
sean-k-mooneyok os it sthe test case that is testign the  cross treadh sharing which is something i explictly did not want to supprot17:07
sean-k-mooneywell i need to look at the test code to see what that is testing17:08
gibisean-k-mooney: thanks for looking. I'm still trying to catch up after my PTO17:09
sean-k-mooneyhttps://github.com/openstack/nova/blob/master/nova/tests/unit/test_utils.py#L1893-L1917 ya it is17:09
sean-k-mooneygibi: no wories did you enjoy the break17:10
sean-k-mooneyi have not tested that in threaded mode so ill run that in a loop and see if it repoduces locally17:10
gibithere were parts I enjoyed very much :)17:10
sean-k-mooneyok ya so that test is not stable17:15
sean-k-mooneyit cant work relyably because i expclity detect the case wehre we context switch betwen thread 1 and thread 2. https://github.com/openstack/nova/commit/22012360c40045ac1bf6dd0dc95aea3e15fed0cd#diff-d6d688e548906bcae5669480a03749c0a52f32634762e23adfbcd8de88b9d980R1236-R124017:15
sean-k-mooneybiscly the exectpiotn that is beign raised happens if the first worker thread enters the context manager while the other worker thread is also in the context manger17:16
sean-k-mooneyif the kernel runs them serially the test will pass17:16
sean-k-mooneyif it decied to interleave them it wont17:16
sean-k-mooneywe have 3 options. i can remove the raise which was orginally there to prevent this type of sharing17:17
sean-k-mooneyi can remove the test17:17
sean-k-mooneyor i can alter the test to prevent the interleving but that not very useful17:18
sean-k-mooneysince the interation of the patch where i added that raise i added a buch of addtion lockign to make it safe to share the lockgrarud between threads17:18
sean-k-mooneyso removing the raise is proably the correct thing to do and just make it a debug log or remove that if and the self._active varible entirly17:19
sean-k-mooneythat raise only exists to provide a singal to use that 17:21
sean-k-mooney        thread1 = threading.Thread(target=worker, args=(1, lock_guard))17:21
sean-k-mooney thread2 = threading.Thread(target=worker, args=(2, lock_guard))17:21
sean-k-mooneyis a programing logic error17:21
sean-k-mooneywith the extra lockign i have added that shoudl now be safe even if i dislike it so the extra protection is not required17:22
sean-k-mooneyany prefernce on how to fix it17:22
sean-k-mooneywe can just say the comments are enouch to discourge that useage and delete the check, we can add a debug log or i can remove the test.17:23
gibiI don't know at the moment. Sorry.17:23
gibiI need to get back to this17:23
sean-k-mooney its fine ill create a quick patch to remvoe the raise and ill ask this quetion on the review17:23
sean-k-mooneyand you can leave feedback when you have time17:23
gibicool. Thanks!17:24
sean-k-mooneywas there a bug for this if not ill just do "Related-Bug: #2048837" to tie it to the orgianal bug fix17:24
sean-k-mooneytkajinam: ^17:25
opendevreviewTakashi Kajinami proposed openstack/nova master: libvirt: Add capability to load loader and nvram from xml  https://review.opendev.org/c/openstack/nova/+/96908617:46
opendevreviewTakashi Kajinami proposed openstack/nova master: libvirt: Add capability to load ssm feature from existing xml  https://review.opendev.org/c/openstack/nova/+/96913118:17
-opendevstatus- NOTICE: An update to one of our base jobs roles broke another base job role. This update has been reverted and jobs should be working again.20:59
JayFgibi: finally got back around to this post-holidays; no code change was needed and I have an explanation: https://review.opendev.org/c/openstack/nova/+/96932122:59

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!