| *** mhen_ is now known as mhen | 01:13 | |
| opendevreview | Merged openstack/devstack-plugin-ceph master: Deploy rbd-mirror daemon https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/948768 | 11:48 |
|---|---|---|
| *** haleyb is now known as haleyb|out | 14:17 | |
| sean-k-mooney | gmaan: have you seen https://paste.opendev.org/show/bp8QWbT7C1hhNA04janf/ before | 16:20 |
| sean-k-mooney | its form a job that im hacking to run with pytho3.13 using pyvenv | 16:20 |
| sean-k-mooney | i dont see how that could possibly be related | 16:21 |
| sean-k-mooney | but have we seen any mysql passowrd setup issues on noble like this? | 16:21 |
| sean-k-mooney | i jsut kicked off a recheck to see if this will repeat | 16:22 |
| sean-k-mooney | im wondering if this is possible related to bindep | 16:23 |
| sean-k-mooney | i used ensure bindep to install gcc ectra so that may have also installed mysql adn caused some issues | 16:24 |
| sean-k-mooney | im checkign the actuly output.txt now but just tought i would ask here to make sure this is specific to my job and not somethign affecting gate jobs in general | 16:25 |
| sean-k-mooney | im going to run with that in the absence of other evidecne and see if i can get this to work | 16:27 |
| sean-k-mooney | my overall goal is to get some more useful data about just how broken nova is on 3.13 due to the eventlet porblems | 16:27 |
| sean-k-mooney | ok the job is as far as installing neutron so it might have already got past where it was failing before | 16:56 |
| clarkb | sean-k-mooney: I think this is unlikely given what I know of devstack setup but in the zuul test suite we discovered that mariabd (not mysql) has a race condition with flush privileges that prevents concurrent logins from occuring | 17:03 |
| clarkb | sean-k-mooney: we fixed that by removing unnecessary flush privileges calls | 17:03 |
| clarkb | upstream also has a fix if you wait for a new mariadb or use mysql | 17:03 |
| sean-k-mooney | ok well it seams it was caused by installign it with bindep | 17:05 |
| sean-k-mooney | i just split out the compile deps into a diffent target and stopped usign test | 17:05 |
| sean-k-mooney | btu then it faisl in nova-manage | 17:05 |
| sean-k-mooney | so my job is now workign and as we kind of expect nova is explodign on python 3.13 because i assume eventlet | 17:06 |
| sean-k-mooney | but now tha ti have somethign kind of workign i can put this on top fo gibi's evenlet series | 17:07 |
| sean-k-mooney | well or too the side | 17:07 |
| sean-k-mooney | via depend-on | 17:07 |
| gmaan | sean-k-mooney: no, I have not seen this before | 17:08 |
| sean-k-mooney | oh no im wrong | 17:08 |
| sean-k-mooney | it failed the same way but later we call nova-mange in post and that also failes | 17:09 |
| sean-k-mooney | clarkb: so sorry you were saying there is a race | 17:09 |
| sean-k-mooney | clarkb: is that patch still up for review | 17:09 |
| sean-k-mooney | i can try pullling it in | 17:09 |
| sean-k-mooney | this is how its failing https://zuul.opendev.org/t/openstack/build/dd9da3eecd304f9a9cb4df57a0a3557e/log/job-output.txt#12013 | 17:10 |
| clarkb | sean-k-mooney: it was in the zuul test suite I don't think its anything to pull in. More just calling out the behavior we discovered since it affects logins to the db which is what your paste indicated problems with | 17:10 |
| clarkb | basically if in the db you run flush privileges there is a small window of time where all new connectiosn will fail to authenticate | 17:10 |
| sean-k-mooney | ack | 17:10 |
| sean-k-mooney | ill need to check the code but i wonder if this is runnign as an asycn job in devstack | 17:13 |
| sean-k-mooney | that would be a no | 17:17 |
| sean-k-mooney | its failing between setign teh passwrod with mysladmin and then trying to create the database user | 17:18 |
| sean-k-mooney | https://github.com/openstack/devstack/blob/master/lib/databases/mysql#L110-L136 | 17:18 |
| sean-k-mooney | ok well i guess i should just tyr to replciate this locally. there sinet much point in wastign ci resouce tryign to debug this this evning | 17:20 |
| sean-k-mooney | i think ill call it a day there | 17:20 |
| sean-k-mooney | enjoy your weekends o/ | 17:20 |
| opendevreview | Fernando Ferraz proposed openstack/devstack-plugin-nfs master: [DNM] Run Glance with Cinder as backend https://review.opendev.org/c/openstack/devstack-plugin-nfs/+/940482 | 18:11 |
| opendevreview | Fernando Ferraz proposed openstack/devstack-plugin-nfs master: [DNM] Run Glance with Cinder as backend https://review.opendev.org/c/openstack/devstack-plugin-nfs/+/940482 | 18:13 |
| opendevreview | Fernando Ferraz proposed openstack/devstack-plugin-nfs master: [DNM] Testing devstack-plugin-nfs success rate https://review.opendev.org/c/openstack/devstack-plugin-nfs/+/951998 | 20:59 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!