14:02:23 <jokke> #startmeeting glance
14:02:24 <openstack> Meeting started Thu Jan 14 14:02:23 2021 UTC and is due to finish in 60 minutes.  The chair is jokke. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:28 <openstack> The meeting name has been set to 'glance'
14:02:31 <jokke> #topic roll-call
14:02:39 <jokke> o/
14:02:57 <jokke> We have very stubby agenda for today
14:03:08 <jokke> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:03:32 <Steap> haha
14:04:04 <jokke> rajivmucheli: if you wait couple of minutes and we'll get to your question in the open discussion
14:04:13 <rajivmucheli> sure
14:04:38 <jokke> #topic release updates
14:04:57 <jokke> Just a reminder milestone 2 is next week
14:05:58 <jokke> Next week is also deadline for specs, so anything still under discussion should be ready by Wed EOB latest with core acks so Abhishek can merge them when he comes back
14:06:38 <jokke> #topic open discussion
14:06:47 <jokke> rajivmucheli: stage is yours
14:07:56 <rajivmucheli> i am planning to upgrade my production from train to victoria release. but i see :
14:07:56 <rajivmucheli> root@glance-78fc69d7d-t4szd:/# glance-api --version
14:07:57 <rajivmucheli> 19.0.0
14:08:13 <rajivmucheli> is this as expected or should it be 21 ?
14:09:05 <jokke> pretty sure it should
14:09:47 <jokke> so Train is 19 Victoria is 21
14:10:11 <rajivmucheli> yes, when i tried upgrading to ussuri, i get 19 instead of 20.
14:10:21 <rajivmucheli> i never experienced this issue till now.
14:10:50 <rajivmucheli> but the site-packages show 21. i wonder why ?
14:11:15 <jokke> hmm-m, what are you using as source? Are you just pulling the sources and installing yourself or are you using some packaged version?
14:11:42 <rajivmucheli> i use loci, enhanced version of kolla.
14:11:57 <rajivmucheli> https://github.com/sapcc/loci
14:13:16 <Steap> python -c "import glance; print(glance.__file__)" <- does this show a weird path or what you'd actually expect?
14:14:16 <rajivmucheli> root@glance-78fc69d7d-t4szd:/var/lib/openstack/bin# python -c "import glance; print(glance.__file__)"
14:14:31 <rajivmucheli> root@glance-78fc69d7d-t4szd:/var/lib/openstack/bin# python -c "import glance; print(glance.__file__)"
14:15:02 <rajivmucheli> idk if the above output is seen, the output for the python cmd is /var/lib/openstack/lib/python3.8/site-packages/glance/__init__.py
14:16:09 <jokke> So either we have something broken in pbr or loci does not update the "bin/" entry or what ever you're using as your path to start it and you actually get still the old version
14:17:27 <rajivmucheli> i have been using loci since 3years and never had this error.
14:18:12 <Steap> do you see the same issue with other OpenStack components?
14:18:33 <rajivmucheli> No, i upgraded barbican service. It works well.
14:19:50 <jokke> rajivmucheli: if you can start the API, even listening to another port and check what API minor version it reports
14:20:41 <rajivmucheli> how can i check the minor API version ? the API service is working.
14:20:49 <jokke> if it reports 2.11 as Images API version to the version query, then you're running Victoria and the error is what we're actually reporting as version and no-ne else has noticed it
14:22:38 <rajivmucheli> how do i check it reports 2.11 ? api debug logs doesnt show much
14:22:39 <jokke> if I recall correctly the version negotiation responds to both / and /version endpoints if you send get request to either
14:27:11 <jokke> Anything else? We can continue this outside of the meeting in #openstack-glance channel
14:27:15 <rajivmucheli> the next question is, is it possible for glance to retry an image_upload operation ?
14:28:14 <rajivmucheli> or do i need to enable swift_buffer_on_upload, swift_upload_buffer_dir
14:28:24 <jokke> that depends how it fails. If it's detected we should revert the image back to "Queued" at which point you can redo upload
14:29:36 <jokke> there is no way to resume so if the image state goes back to queued you need to restart the whole upload
14:29:59 <rajivmucheli> i get this error https://bugs.launchpad.net/glance-store/+bug/1673280
14:30:01 <openstack> Launchpad bug 1673280 in glance_store "BackendException: Failed to add object to Swift. - While trying to upload image size of around 400 MB" [Undecided,Incomplete]
14:31:56 <jokke> ohh
14:32:10 <jokke> so that's still a thing
14:32:25 <rajivmucheli> yes.
14:32:54 <jokke> could you put the details into that bug so we can keep tracking it. Last time by the looks of it the original reporter just disappeared
14:33:11 <jokke> so it might be some setting related, but it would be good to get to the bottom of it
14:34:11 <rajivmucheli> sure, can i ping you directly ?
14:34:20 <jokke> sure
14:34:30 <rajivmucheli> thats it from my end. Thanks
14:35:32 <jokke> ok, and like said feel free to join us on #openstack-glance channel for any further digging into the version stuff etc.
14:35:43 <jokke> Steap: you have anything?
14:37:39 <jokke> I take that as no
14:37:45 <jokke> Thanks everyone!
14:37:53 <jokke> #endmeeting