| *** Liang__ has joined #openstack-glance | 01:15 | |
| *** jmlowe has quit IRC | 02:42 | |
| *** wxy has quit IRC | 03:13 | |
| *** ratailor has joined #openstack-glance | 04:28 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-glance | 04:33 | |
| *** gyee has quit IRC | 04:48 | |
| *** udesale has joined #openstack-glance | 05:24 | |
| *** tkajinam has quit IRC | 05:34 | |
| *** tkajinam has joined #openstack-glance | 05:35 | |
| *** tkajinam has quit IRC | 05:50 | |
| *** tkajinam has joined #openstack-glance | 05:51 | |
| abhishekk | dansmith, ack | 05:51 |
|---|---|---|
| openstackgerrit | Merged openstack/glance-specs master: Calculate virtual size of image https://review.opendev.org/741121 | 06:01 |
| openstackgerrit | Merged openstack/glance stable/ussuri: Fix active image when all uploads fail https://review.opendev.org/744997 | 06:26 |
| *** bhagyashris is now known as bhagyashris|off | 07:13 | |
| *** k_mouza has joined #openstack-glance | 07:27 | |
| *** k_mouza has quit IRC | 07:27 | |
| *** Liang__ has quit IRC | 07:51 | |
| *** ralonsoh has joined #openstack-glance | 07:54 | |
| *** Liang__ has joined #openstack-glance | 08:00 | |
| *** ratailor has quit IRC | 08:21 | |
| *** rosmaita has quit IRC | 08:37 | |
| openstackgerrit | Merged openstack/glance master: Stream-friendly disk format inspection module https://review.opendev.org/744205 | 08:48 |
| *** rosmaita has joined #openstack-glance | 08:49 | |
| *** tkajinam has quit IRC | 09:17 | |
| *** ratailor has joined #openstack-glance | 09:24 | |
| *** ratailor_ has joined #openstack-glance | 09:32 | |
| *** k_mouza has joined #openstack-glance | 09:33 | |
| *** ratailor has quit IRC | 09:35 | |
| *** k_mouza has quit IRC | 10:43 | |
| openstackgerrit | Abhishek Kekane proposed openstack/glance_store master: PoC Cinder multiple stores https://review.opendev.org/742371 | 10:45 |
| *** k_mouza has joined #openstack-glance | 10:51 | |
| *** Liang__ has quit IRC | 10:58 | |
| *** k_mouza has quit IRC | 11:46 | |
| *** k_mouza has joined #openstack-glance | 11:56 | |
| *** k_mouza has quit IRC | 12:00 | |
| *** rcernin has quit IRC | 12:32 | |
| *** ratailor_ has quit IRC | 12:43 | |
| *** udesale_ has joined #openstack-glance | 12:44 | |
| *** udesale has quit IRC | 12:47 | |
| *** takamatsu has quit IRC | 12:52 | |
| *** jmlowe has joined #openstack-glance | 13:00 | |
| *** dirk has quit IRC | 13:31 | |
| *** rcernin has joined #openstack-glance | 13:32 | |
| *** rcernin has quit IRC | 13:37 | |
| openstackgerrit | Merged openstack/glance_store master: requirements: Drop os-testr https://review.opendev.org/740226 | 13:50 |
| *** k_mouza has joined #openstack-glance | 13:56 | |
| *** k_mouza has quit IRC | 14:14 | |
| *** k_mouza has joined #openstack-glance | 14:15 | |
| abhishekk | dansmith, jokke if you are around and have some time kindly have a look at https://review.opendev.org/695152 cinder multiple stores specs | 15:01 |
| abhishekk | I need to get it merged today | 15:01 |
| dansmith | sorry, I forgot you require all cores to ack those | 15:02 |
| abhishekk | no problem | 15:03 |
| *** ralonsoh has quit IRC | 15:08 | |
| *** ralonsoh has joined #openstack-glance | 15:08 | |
| dansmith | abhishekk: two quick questions in there | 15:09 |
| abhishekk | looking | 15:09 |
| abhishekk | dansmith, replied to your comments, makes sense to break that in two different sections will do it immediately | 15:16 |
| dansmith | abhishekk: okay, sorry to make you rev again, but I think it'll look better | 15:16 |
| abhishekk | agree, not an issue, just wondering I should mention "original ussuri config" or "config before upgrade" and "config after upgrade" | 15:17 |
| abhishekk | dansmith, ^^ | 15:19 |
| dansmith | I think it'd help yeah | 15:19 |
| dansmith | I would probably avoid saying the release name, | 15:24 |
| dansmith | just to make it more timeless | 15:24 |
| dansmith | because someone upgrading from train->W won't have an "ussuri config" | 15:24 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance-specs master: Make cinder driver compatible with multiple stores https://review.opendev.org/695152 | 15:25 |
| abhishekk | dansmith, ^^ | 15:26 |
| abhishekk | will carry forward rosmaita and smcginnis +2 as it's just a cosmetic change | 15:26 |
| dansmith | cool | 15:28 |
| dansmith | abhishekk: have you seen this error before? https://ed75466bb08c2322add8-83eda0bc1935c183e842d53b008f8f00.ssl.cf5.rackcdn.com/743597/12/check/openstack-tox-functional-py36/8bb36d0/testr_results.html | 15:28 |
| dansmith | specifically this: 2020-08-06 18:38:34.524 3668 ERROR glance.async_.taskflow_executor FileNotFoundError: [Errno 2] No such file or directory: '/tmp/tmpxt2cb_bz/images_3/e3debcaf-3c14-4ea4-ba82-2c1c14a8970a' | 15:28 |
| dansmith | it's similar to an error I'm seeing in my test where my database file seems to disappear from underneath me | 15:28 |
| abhishekk | nope | 15:28 |
| dansmith | hrm | 15:29 |
| dansmith | if I run --until-failure for about ten minutes I can reproduce one of the two of those | 15:29 |
| abhishekk | images_3 is one of your filesystem store from multiple stores rigjt | 15:30 |
| dansmith | yes, but this isn't my test, this was here already | 15:32 |
| dansmith | test_copy_image_revert_lifecycle | 15:32 |
| *** rcernin has joined #openstack-glance | 15:33 | |
| abhishekk | dansmith, hmm | 15:34 |
| dansmith | also, I wonder if we could squelch the stevedore.extension logging in the test base class so that we don't get such a huge dump on each failure? | 15:35 |
| dansmith | I mean, I think we can, I wonder if that would be a welcome change :) | 15:35 |
| abhishekk | absolutely | 15:36 |
| *** rcernin has quit IRC | 15:38 | |
| abhishekk | dansmith, I would suggest we should skip this test or remove it and it is flaky and depends on removing one of the configured filestore to produce the failure | 15:38 |
| abhishekk | This test needs to be rewritten more carefully (similar to something how you have written test to bust lock or something like that) | 15:40 |
| dansmith | so, looking at the test, I guess you're trying to delete the file? | 15:40 |
| dansmith | maybe the test is just racing with the task and checks the stores before the task has finished? | 15:40 |
| abhishekk | I am trying to delete the file store directory itself | 15:40 |
| dansmith | right okay | 15:40 |
| dansmith | looks really similar to my failure, but maybe it's not the same | 15:41 |
| abhishekk | yes, when I added it, it was running ok for couple of weeks and suddenly started failing occasionally | 15:42 |
| abhishekk | and there was not a easy way to produce failure in the functional tests than this :( | 15:42 |
| dansmith | oh okay I thought you didn't recognize the fail? or you just didn't recognize that FileNotFoundError part of it? | 15:45 |
| abhishekk | I do recognized it now earlier I replied just based on your error message you pasted here | 15:46 |
| dansmith | gotcha, okay | 15:48 |
| dansmith | this patch hit it too: https://review.opendev.org/#/c/744011/ | 15:48 |
| dansmith | in the last day | 15:48 |
| abhishekk | yeah | 15:49 |
| abhishekk | I think its better to skip/remove that test at it is not full proof | 15:50 |
| dansmith | can we put a retry in there maybe? | 15:50 |
| abhishekk | may be but not sure it will help | 15:52 |
| dansmith | well, if it's a race with the task finishing I would think it would.. I'll try to poke at it a bit when I'm done revising this set | 15:53 |
| abhishekk | ack | 15:55 |
| * abhishekk going for dinner will be back in 45 mins | 16:07 | |
| openstackgerrit | Dan Smith proposed openstack/glance master: Heartbeat the actual work of the task https://review.opendev.org/743426 | 16:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Update task message during import https://review.opendev.org/743427 | 16:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Implement time-limited import locking https://review.opendev.org/743597 | 16:31 |
| *** udesale_ has quit IRC | 16:41 | |
| openstackgerrit | Dan Smith proposed openstack/glance master: Squelch the stevedore.extension debug logging in functional tests https://review.opendev.org/745374 | 16:44 |
| dansmith | abhishekk: ^ | 16:45 |
| abhishekk | dansmith, ack | 16:45 |
| smcginnis | dansmith: Where are those logs happening? I haven't noticed anything in our functional test logs. | 16:47 |
| abhishekk | dansmith, one question about https://review.opendev.org/#/c/743597/13/glance/api/v2/images.py line #122 | 16:47 |
| dansmith | smcginnis: when something starts an api and then the test fails, let me get you a link] | 16:48 |
| dansmith | smcginnis: check this out: https://ed75466bb08c2322add8-83eda0bc1935c183e842d53b008f8f00.ssl.cf5.rackcdn.com/743597/12/check/openstack-tox-functional-py36/8bb36d0/testr_results.html | 16:48 |
| smcginnis | dansmith: Ah, I was only looking at a successful run. That makes sense. | 16:48 |
| smcginnis | Yeah, pretty noisy and useless. | 16:48 |
| dansmith | yeah and when they fail running locally, it blows my terminal buffer all to pieces :) | 16:48 |
| dansmith | abhishekk: shoot | 16:49 |
| abhishekk | Will it be a possibility that someone might have actually deleted the image? | 16:49 |
| abhishekk | the message and error response might be misleading or it will be very very corner case? | 16:49 |
| dansmith | abhishekk: yes, image delete will cause you to hit that if timed right, | 16:51 |
| dansmith | but I think the messages all still make sense.. meaning "we lost the race to delete it" still makes sense if the task lock is gone because the image is gone, I think | 16:51 |
| abhishekk | hmm, I am more doubtful about Conflict response here | 16:52 |
| dansmith | and, the user doesn't know if the 409 is because the image is still locked with a valid lock, which is why we emulate the "lock is still valid" error there, since we didn't get to it before the other thread did ... something with it | 16:52 |
| dansmith | a user that gets a 409 just needs to know "this isn't ready yet", and if someone deletes the whole image right after that (or seemingly at the same time), or grabs the lock because it expired and they grabbed it first, it's the same thing from the client's perspective | 16:54 |
| abhishekk | ok | 16:54 |
| dansmith | I think the important thing is to make the "what do I do if I get this error" behavior the same for the user, regardless of the actual reasoning, which is why I raise 409 there as well | 16:55 |
| dansmith | 409 means "you can't have this right now, check back later" | 16:55 |
| abhishekk | yeah, I was just thinking that after sometime I will get 404 :D | 16:55 |
| dansmith | the image was valid when we entered POST /import, so we don't need to return a 404 I think | 16:55 |
| abhishekk | yes, makes sense | 16:56 |
| dansmith | yep, but you can always get a 404 pretty much any time I think | 16:56 |
| abhishekk | May be I was thinking we should add "or Image is deleted between the process" to the warning message but its ok IMO | 16:57 |
| dansmith | well, we can't really differentiate between the image being gone and the lock being gone atomically without a lot of work, and if we checked a second time, we might report the wrong thing because it was deleted between the two checks | 16:59 |
| abhishekk | yep, makes sense | 17:00 |
| *** ralonsoh has quit IRC | 17:02 | |
| dansmith | abhishekk: ahh, I see that copy_image_revert_lifecycle test is trying to delete the images_3 during the task execution before it gets that far to trigger a failure, right? | 17:04 |
| dansmith | so it's kinda racy by definition | 17:05 |
| abhishekk | dansmith, yes | 17:05 |
| dansmith | abhishekk: can't we do the delete before we even initiate the import? then it will succeed on the first store, and when _ImportToStore('images_3 | 17:05 |
| dansmith | ') runs it will fail? | 17:05 |
| dansmith | that will be less racy I think | 17:05 |
| abhishekk | just a min let me have a look at that again | 17:06 |
| *** k_mouza has quit IRC | 17:06 | |
| dansmith | seems to work locally | 17:07 |
| dansmith | maybe I will push up that change and we can recheck it a bunch of times | 17:07 |
| abhishekk | yes I think we should delete it earlier | 17:07 |
| abhishekk | even after the second line of test it self where we actually starting the services | 17:07 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Fix non-deterministic copy_image_revert_lifecycle test https://review.opendev.org/745379 | 17:09 |
| dansmith | well, I think it's better right before the import so it's clear why we're doing it | 17:10 |
| dansmith | but let's see if we can get that ^ to fail | 17:10 |
| abhishekk | ack | 17:10 |
| dansmith | I guess we might as well just merge that since it works, and then we'll get all the testing of everything else running on top of master | 17:11 |
| abhishekk | makes sense | 17:13 |
| abhishekk | dansmith, also adding data_files in glance_store/setup.cfg sounds more reasonable | 17:17 |
| dansmith | if people will need that for real, then... definitely :) | 17:17 |
| abhishekk | yes they will | 17:18 |
| dansmith | I'm super glad you're working on a job to make that all get tested.. in addition to just proving that it works, it helps uncover things like this too | 17:19 |
| abhishekk | ++ | 17:20 |
| *** rcernin has joined #openstack-glance | 17:34 | |
| *** rcernin has quit IRC | 17:39 | |
| *** k_mouza has joined #openstack-glance | 17:44 | |
| openstackgerrit | Merged openstack/glance-specs master: Make cinder driver compatible with multiple stores https://review.opendev.org/695152 | 17:48 |
| *** k_mouza has quit IRC | 17:49 | |
| dansmith | abhishekk: I think it's safe to merge that determinism change and let the full stream run on it | 18:04 |
| abhishekk | dansmith, sounds good | 18:05 |
| abhishekk | I have also ran it locally multiple times and it is working (hope it will work in gates as well :D) | 18:10 |
| dansmith | yeah | 18:13 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Add functional test for task status updating https://review.opendev.org/745392 | 18:58 |
| dansmith | dang this database file thing | 19:24 |
| abhishekk | 1 minute to merge | 19:30 |
| openstackgerrit | Merged openstack/glance master: Squelch the stevedore.extension debug logging in functional tests https://review.opendev.org/745374 | 19:30 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Add testing for _CompleteTask in api_image_import https://review.opendev.org/743593 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Add tests for _ImportToStore.execute() https://review.opendev.org/743595 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Flesh out FakeImage for extra_properties https://review.opendev.org/743839 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Add image_delete_property_atomic() helper https://review.opendev.org/743596 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Heartbeat the actual work of the task https://review.opendev.org/743426 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Update task message during import https://review.opendev.org/743427 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Implement time-limited import locking https://review.opendev.org/743597 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: Add functional test for task status updating https://review.opendev.org/745392 | 19:31 |
| openstackgerrit | Dan Smith proposed openstack/glance master: DNM: Try to debug spurious DB file failure https://review.opendev.org/745398 | 19:31 |
| openstackgerrit | Merged openstack/glance master: Fix non-deterministic copy_image_revert_lifecycle test https://review.opendev.org/745379 | 19:38 |
| *** k_mouza has joined #openstack-glance | 19:45 | |
| *** k_mouza has quit IRC | 19:50 | |
| openstackgerrit | Abhishek Kekane proposed openstack/glance_store master: Copy data files to glance upon installation https://review.opendev.org/745407 | 19:54 |
| abhishekk | dansmith, ^^ | 19:55 |
| * abhishekk signing out for the day, happy weekend all | 19:56 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/python-glanceclient master: [goal] Migrate testing to ubuntu focal https://review.opendev.org/744075 | 21:30 |
| openstackgerrit | Ghanshyam Mann proposed openstack/python-glanceclient master: Update test certificates to use strong signing https://review.opendev.org/718093 | 21:31 |
| openstackgerrit | Ghanshyam Mann proposed openstack/python-glanceclient master: [goal] Migrate testing to ubuntu focal https://review.opendev.org/744075 | 21:31 |
| *** rcernin has joined #openstack-glance | 21:36 | |
| *** rcernin has quit IRC | 21:41 | |
| *** rcernin has joined #openstack-glance | 21:47 | |
| *** rcernin has quit IRC | 22:26 | |
| *** rcernin has joined #openstack-glance | 22:31 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/python-glanceclient master: [goal] Migrate testing to ubuntu focal https://review.opendev.org/744075 | 22:37 |
| *** k_mouza has joined #openstack-glance | 23:10 | |
| *** k_mouza has quit IRC | 23:14 | |
| *** rcernin has quit IRC | 23:38 | |
| *** rcernin has joined #openstack-glance | 23:59 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!