| opendevreview | Joan Gilabert proposed openstack/watcher master: Change id field to uuid when appropiate https://review.opendev.org/c/openstack/watcher/+/975200 | 08:28 |
|---|---|---|
| opendevreview | Joan Gilabert proposed openstack/watcher master: Use wrapper classes for novaclient objects https://review.opendev.org/c/openstack/watcher/+/972913 | 08:28 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 08:28 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 08:28 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 08:28 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 08:28 |
| opendevreview | chandan kumar proposed openstack/watcher-dashboard master: [poc]Add Playwright-based E2E testing framework https://review.opendev.org/c/openstack/watcher-dashboard/+/970353 | 09:33 |
| dviroel | sean-k-mooney: ack, it fixes because it retries the operation in the end. The possible solutions are retrying or serialize. But the lock in database is still there. Good catch on the missing decorator | 10:53 |
| sean-k-mooney | yep that is exactly why its passing. i spoke to steph breifly and i am goign to propose modifying oslo.db i just is unlikely to happen this cycle | 11:08 |
| sean-k-mooney | so have oslo.db nativly have that excption handeler to convert database locked to a dbdeadlock excption instead of operational error | 11:08 |
| sean-k-mooney | the non clinet lib freeze and therefor oslo.db freeze is next week so im going to look at doign that for next cycle | 11:09 |
| sean-k-mooney | it will help projects like ironic who supprot sqlite in production | 11:10 |
| sean-k-mooney | one thing i need to follow up with is checkign that we are not missing any other decorators | 11:10 |
| sean-k-mooney | dviroel: are you ok with that overall approch. i tried seraisliging the writes as well but that ended up with deadlocks in the tests | 11:11 |
| sean-k-mooney | hence the pivot to retry | 11:12 |
| dviroel | i'm ok with retrying but i don't like the way that we touch oslo.db internals and reuse the deadlock retry for that scenario | 11:14 |
| dviroel | do you plan to propose changing the deadlock retry in oslo, or the isolation level that is mentioned in the commit msg? | 11:14 |
| amoralej_ | so, adding the decorator fixes the ci? | 11:14 |
| dviroel | amoralej_: no, it also need to modify de retry filter | 11:15 |
| dviroel | the deadlock retry search for different errors | 11:15 |
| amoralej_ | ah, got it | 11:15 |
| dviroel | btw, the missing retry_on_deadlock could even be backported as bug | 11:16 |
| dviroel | in the update operatrion ^ | 11:16 |
| dviroel | sean-k-mooney: I tried with claude yesterday, and serializing was working. Let me push the change so we can have more options | 11:17 |
| amoralej_ | wrt the osprofiler patch, https://review.opendev.org/c/openstack/watcher/+/976280 i used a similar approach as other projects of wrapping the osprofiler decorators to ignore if the osprofiler is not installed or conf.profiler.enabled is false in https://review.opendev.org/c/openstack/watcher/+/976280/1/watcher/common/profiler.py. | 11:26 |
| amoralej_ | However, to make that work in all cases we need to make sure that the traced classes are loaded after the configuration. The only way i was able to make it work was moving the import in https://review.opendev.org/c/openstack/watcher/+/976280/1/watcher/cmd/eventlet/applier.py and https://review.opendev.org/c/openstack/watcher/+/976280/1/watcher/cmd/eventlet/decisionengine.py but i feel there must be a better approach, any idea about a | 11:27 |
| amoralej_ | better alternative? | 11:27 |
| rlandy | jgilaber: hi - could I get your review on https://review.opendev.org/c/openstack/releases/+/976016? I marked it W-1 so nobody merges it w/o watcher core approval | 11:29 |
| opendevreview | Merged openstack/watcher master: Deprecate MAAS integration https://review.opendev.org/c/openstack/watcher/+/974308 | 11:30 |
| dviroel | amoralej_: i will take a look in a minute | 11:34 |
| amoralej_ | sure, thanks! | 11:34 |
| jgilaber | rlandy, sure, I'll review it shortly, already started this morning | 11:40 |
| rlandy | thank you - no rush | 11:41 |
| opendevreview | Douglas Viroel proposed openstack/watcher master: Enable Applier parallel engine in native thread mode https://review.opendev.org/c/openstack/watcher/+/975522 | 11:51 |
| sean-k-mooney | dviroel: yes i plan to propsoe that to oslo.db sepreatly | 11:57 |
| sean-k-mooney | that does not mean we cant proceed indepently with the serialzed approch | 11:58 |
| dviroel | right, I would like a fix for watcher within this cycle, and we can revert it once we have a fix available in oslo.db | 12:00 |
| sean-k-mooney | dviroel: funny i tried something very clsoe to https://review.opendev.org/c/openstack/watcher/+/975522/3/watcher/tests/fixtures/db_lock.py first hut i didnt have a timeout on the lock. | 12:00 |
| dviroel | sean-k-mooney: that was claude's suggestion | 12:01 |
| sean-k-mooney | https://review.opendev.org/c/openstack/watcher/+/975522/3/watcher/tests/unit/db/base.py has a bug bvy the way you are leakyign the tempory dbs | 12:01 |
| sean-k-mooney | you shoudl either use the tempder fixture or you need to self.addCleanup on the dbfile | 12:01 |
| sean-k-mooney | https://github.com/testing-cabal/fixtures/blob/master/fixtures/_fixtures/tempdir.py | 12:02 |
| sean-k-mooney | https://github.com/testing-cabal/fixtures/blob/master/fixtures/_fixtures/mockpatch.py#L35-L45 | 12:02 |
| sean-k-mooney | dviroel: let me seperate out a clean patch to add the missing decorators | 12:03 |
| dviroel | +1 | 12:03 |
| sean-k-mooney | if you can refine the tempfile handelign we cna likey proceed with your version and i can follow up with oslo db changes without rushing | 12:03 |
| dviroel | sean-k-mooney: sure, I will work on it a bit more | 12:04 |
| dviroel | lets see if we have also a green CI | 12:04 |
| dviroel | * works on my machine * | 12:05 |
| sean-k-mooney | oh the luxury | 12:05 |
| sean-k-mooney | a ci that is green is often overrated | 12:06 |
| dviroel | indeed | 12:06 |
| sean-k-mooney | *underated | 12:07 |
| sean-k-mooney | over looked | 12:07 |
| sean-k-mooney | i dong know whch which word i was going for but i need to go chekc if the all the devstack relate patch for pkg_resouce are in | 12:08 |
| dviroel | ++ | 12:08 |
| dviroel | amoralej_: I need more time to think about, this approach is a bit different from what I already see in manila I think | 12:34 |
| *** haleyb_ is now known as haleyb | 13:36 | |
| amoralej_ | dviroel, lemme check manila, i checked nova and glance | 13:40 |
| amoralej_ | yeah, manilla follows a totally different approach ... | 13:47 |
| opendevreview | Alfredo Moralejo proposed openstack/watcher master: Skip volume_migrate actions in pre_condition phase https://review.opendev.org/c/openstack/watcher/+/976393 | 14:05 |
| amoralej_ | but that restricts the options to specify the tracing points, as it allows only to enable osprofiler traces in the service manager classes or in https://github.com/openstack/manila/blob/22f33460e1c5b127cce7ff747f1de307d7c5fa89/manila/service.py#L121 or the shares driver classes in https://github.com/openstack/manila/blob/22f33460e1c5b127cce7ff747f1de307d7c5fa89/manila/share/manager.py#L333 not any arbitrary class | 14:22 |
| opendevreview | Alfredo Moralejo proposed openstack/watcher master: Add OSProfiler support for cross-service tracing https://review.opendev.org/c/openstack/watcher/+/976280 | 15:07 |
| opendevreview | Alfredo Moralejo proposed openstack/watcher master: Skip volume_migrate actions in pre_condition phase https://review.opendev.org/c/openstack/watcher/+/976393 | 15:16 |
| opendevreview | Douglas Viroel proposed openstack/watcher master: Enable Applier parallel engine in native thread mode https://review.opendev.org/c/openstack/watcher/+/975522 | 15:52 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Use wrapper classes for novaclient objects https://review.opendev.org/c/openstack/watcher/+/972913 | 18:07 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 18:07 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 18:07 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 18:07 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 18:07 |
| opendevreview | Merged openstack/watcher master: Fix term.py with 0.22.2 https://review.opendev.org/c/openstack/watcher/+/964762 | 19:11 |
| opendevreview | Merged openstack/watcher master: devstack: Drop ineffective port variables https://review.opendev.org/c/openstack/watcher/+/974647 | 21:01 |
| opendevreview | Merged openstack/watcher master: Deprecate unused [DEFAULT] bindir https://review.opendev.org/c/openstack/watcher/+/975808 | 21:01 |
| opendevreview | Merged openstack/watcher master: Remove eventlet-based timeout in CDM collectors https://review.opendev.org/c/openstack/watcher/+/968568 | 21:01 |
| opendevreview | Merged openstack/watcher master: Drop redundant N366 check https://review.opendev.org/c/openstack/watcher/+/972300 | 21:01 |
| opendevreview | Merged openstack/watcher master: Remove old checks for Python 2-3 compatibility https://review.opendev.org/c/openstack/watcher/+/972289 | 21:01 |
| opendevreview | Merged openstack/watcher master: Remove url tags from README https://review.opendev.org/c/openstack/watcher/+/970339 | 21:01 |
| opendevreview | Merged openstack/watcher stable/2025.2: Make VM migrations timeout configurable and apply reasonable defaults https://review.opendev.org/c/openstack/watcher/+/969483 | 22:54 |
| opendevreview | Merged openstack/watcher master: devstack: Remove ineffective ssl options https://review.opendev.org/c/openstack/watcher/+/975809 | 23:03 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!