Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2025-09-05
14:52:40 sean-k-mooney like i had it open the entire tiem and was workign on it on an doff since i said i would on irc
14:52:56 sean-k-mooney i kept stopping to explian how to fix it to bauzas
14:53:20 dansmith sean-k-mooney: that seems like a rather uncharitable characterization of the conversation as I remember it
14:54:18 bauzas okay, I don't want to create a situation, lemme abandon my change
14:54:30 bauzas this is Friday afternoon for me and I'm done arguing for the week :)
14:54:45 sean-k-mooney if your change is done and has all the test and what gibi asked for i litrally only have one test left ot finsh
14:55:01 sean-k-mooney so i can drop it but i was almost done before my team call earier today
14:55:25 bauzas I don't f... care :)
14:55:33 bauzas if you have time to work on it, do it
14:56:09 bauzas anyway, patch abandoned, floor is yours but please make sure that's on a good shape for rc1
14:56:23 bauzas I can't just argue, again I'm done
14:56:25 dansmith I don't want to go dig up actual quotes, but it surely seemed to me we left the conversation with sean-k-mooney indicating not a lot of time to work on it and bauzas saying he did
14:56:51 dansmith so I reviewed bauzas' immediately and then sean-k-mooney pushed a competing one up after even that first round of reviews
14:56:53 bauzas fwiw, we have a tracker somewhere downstream that also needs to reflect the who does what
14:57:01 dansmith so now I've got to go review the other one?
14:57:04 bauzas so please update it as well
14:57:39 bauzas dansmith: the two are basically identical, except an object modification (which hopefully is just an helper)
14:58:00 bauzas tests diverge a little bit
14:58:07 dansmith so even less necessary to have pushed up a competing patch after the first one was the same solution?
14:58:31 bauzas that's 5pm here on a Friday anyway and I started at 8am, I call it a wrap
15:01:28 sean-k-mooney so i honestly done know what to say
15:01:56 sean-k-mooney bauzas: almost made me rage quite since it feld liek tehy were inteally not reading anythin i wrote on irc
15:01:58 bauzas well, my patch is abandoned, the problem is over
15:02:23 sean-k-mooney so i stop what i was workign on and said i woudl owrk on it
15:02:53 sean-k-mooney i spemnd my evning na dmoning workign on it between interups and while i was waiting for tox runs you pushed a pathc
15:03:36 bauzas that's my point, I had capacity, you didn't had, which is why I offered to push it
15:03:42 sean-k-mooney that patch didnt have hte fallback or tests which i had in my locally so i contineus working on it
15:03:53 bauzas now you say that you have capacity, which is cool, then I abandon mine
15:04:25 bauzas if that sounds a matter of IP, then surely I don't give a single cent about it
15:04:31 sean-k-mooney no i made time to work on it because it felt like you were ignoring me
15:04:44 sean-k-mooney you almost made me rage quite irc yesterday
15:05:09 bauzas the good news is that irc logs are public
15:05:43 bauzas I don't think I had the intent to make you rage, I just expressed one thing about capacity and again I offered a solution
15:06:06 bauzas now, again, fact is that my patch is ABANDONED, so we anyway need to review your patch
15:06:07 opendevreview Pavlo Shchelokovskyy proposed openstack/nova master: Fix num_instances_weight_multiplier option example https://review.opendev.org/c/openstack/nova/+/959866
15:06:30 sean-k-mooney bauzas: im currently reviewing your
15:06:37 bauzas please then update yours based on gibi's feedback and then I'll review it
15:06:38 sean-k-mooney to see what the delta is
15:07:29 sean-k-mooney i have, im litrally just workign ont hte unit test to show the case where the rpc is called and hwen it not based on if the stats have uptime
15:08:08 sean-k-mooney that was the only bit of feedback i didnt have completed 2 hours ago
15:12:48 bauzas sean-k-mooney: I really have to drop soon, again, fact is, I had time today to work on a PS3, no other bits than that
15:13:42 sean-k-mooney bauzas: enjoy your weekend sylvain
15:13:43 bauzas I don't have a particular opinion of my contributions fwiw, so again, feel free to update your patch so I can review it
15:14:30 sean-k-mooney and yes ill finish updating it but i feel like we need to reset
15:15:05 sean-k-mooney we really need ot just chil and hang out some time
15:15:14 dansmith I must say, I feel like something pretty uncool has happened here
15:19:12 sean-k-mooney we both obviously were not comunicating well and both feel like we have wated each other time. syvalisn patch does nto have hte optimisation to prevent the slow down for dirver that dont supprot uptime, that the only delta between the two in general at this point
15:19:46 dansmith sean-k-mooney: I had already commented on that on his patch before you even posted yours, so that's not a reason
15:20:10 dansmith sean-k-mooney: also, he marked you as co-author so if you had tweaks you could have pushed them directly there after it was up, or written tests for his patch instead of writing your own and then tests for that
15:20:20 dansmith so it really feels a lot like a snipe to me, sorry to say
15:21:03 dansmith his PoC had also been posted in the middle of the ending part of the conversation, logged here so it's not like it was a surprise that it was posted or anything
15:22:19 dansmith I even asked you if you had a patch or just a solution "in your head" and you confirmed the latter so.. I think it was pretty fair game at that point, especially with the mention of not having time to work on it (or a sense that it was a priority)
15:22:28 sean-k-mooney it very much was to me because i didn even know he has started on it
15:23:13 dansmith sorry, but that feels dishonest to me.. he posted it and pinged you by name
15:24:18 dansmith and you even responded and said yours was different and "partly done" .. to me that sounds like it _wasn't_ different (at least in the grand sense) and also not like ready to submit
15:24:57 dansmith anyway, my skin in the game is minimal since I had just reviewed one that is not abandoned, but from the somewhat independent observer, this feels highly unfair and unconstructive to me
15:25:57 sean-k-mooney partly done was i was fixign the unit test which sylvain skipped
15:26:14 sean-k-mooney if we want to resotre and merge sylvains we can
15:27:23 sean-k-mooney what im annoyed byt most is listrlly soemthin that i had debugs a few weeks ago, planed how to fix and had plan to do after RC1 was made suddenly really urgent
15:28:44 sean-k-mooney and instead of actully takign the tiem to listen to the backward compatibale solution that we can backport it was repwadly suggested we break our api contract or add config drive api behevaior
15:30:39 dansmith that is not a reasonable characterization of the conversation we had *at all*
15:34:29 sean-k-mooney did you actully look at where the converstaion started 3-4 hours before you came on line
15:34:56 dansmith yes
15:35:37 opendevreview Elod Illes proposed openstack/osc-placement master: [CI] Drop unsupported py39 based functional job https://review.opendev.org/c/openstack/osc-placement/+/959873
15:36:15 opendevreview Merged openstack/nova master: [PCI tracker]Remove non configured devs when freed https://review.opendev.org/c/openstack/nova/+/954613
15:36:22 opendevreview Elod Illes proposed openstack/osc-placement stable/2025.2: [CI] Drop unsupported py39 based functional job https://review.opendev.org/c/openstack/osc-placement/+/959874
15:36:32 opendevreview Merged openstack/nova master: [pci]Keep used dev in Placement regardless of dev_spec https://review.opendev.org/c/openstack/nova/+/954149
15:38:15 sean-k-mooney i lirtally startded workin on it when i said "bauzas: look im just going to go fix this becuase im -2 on just riping it ou8t" jkulik provided me to the link to the bug gibi filed earlier and i checcked out the branch adn started workign on it
15:40:06 sean-k-mooney every time i replied on irc or slack after that yesterdy until i push my patch i was swappign form it to reply
15:53:47 bauzas I just saw it from the logs, I have to say
15:54:14 bauzas after that I said that I'd like to have a consensus about the solution, and then I moved it quickly
15:54:57 bauzas so I understand your concern, but please understand as well that my main point is that I want that fix to be part of RC1 and backportable, now that we have a consensus
15:55:02 sean-k-mooney i think that were we diverged
15:55:11 sean-k-mooney the was part is my unit test is stil failing...
15:55:51 bauzas sean-k-mooney: if you look at the API test, I had to modify some knob given we're directly passing back the object
15:56:13 bauzas you need to delete the stats field before returning the whole dict
15:56:31 sean-k-mooney right os i was tryitng to avoid doing exacty what you did
15:56:49 bauzas (that's the only piece where I used Cursor, in order to understand the root cause but I wrote the code myself)
15:57:11 sean-k-mooney you added https://review.opendev.org/c/openstack/nova/+/959571/3/nova/tests/unit/api/openstack/compute/test_hypervisors.py#80
15:57:19 sean-k-mooney which woudl work and i coudl od it too
15:57:47 bauzas on the second compute
15:57:51 bauzas not on the first one
15:57:51 sean-k-mooney i have been tryign to not change that but fake_compute_node_statistics
15:58:16 bauzas we have two fake computes, I updated the second one and then I skim the stats field before passing the object back
15:58:34 bauzas anyway, 6pm here on a Friday, /me stops
15:58:41 sean-k-mooney ya so i way trying to avoid that because i dint want to change teh v21 data
15:58:53 sean-k-mooney but i think ill just do what you did instead
15:59:10 sean-k-mooney we are using self.stub_out('nova.db.main.api.compute_node_statistics',
15:59:12 sean-k-mooney fake_compute_node_statistics)
15:59:33 sean-k-mooney and i was try8ign to figure out how to refacto rthat so i coudl update the mock only in one specific test
16:00:13 sean-k-mooney im making it a lot harder for myself by doing that
16:01:35 sean-k-mooney effectily i was tryign to avoid touchign the test setup for the older micoversion but you approch is not matirally changing what they are testegin so ill adopt that instead
16:08:48 sean-k-mooney oh i see your testing show not list. ill do both. and include your test in my version
16:32:26 opendevreview Merged openstack/nova master: libvirt: Disable VMCoreInfo device for SEV-encrypted instances https://review.opendev.org/c/openstack/nova/+/958868
16:32:50 opendevreview Merged openstack/nova stable/2024.2: Reproduce that only half of the PCI devs are removed https://review.opendev.org/c/openstack/nova/+/959310
16:33:00 opendevreview sean mooney proposed openstack/nova master: hypervisors: Optimize uptime retrieval for better performance https://review.opendev.org/c/openstack/nova/+/959604
17:36:36 opendevreview Merged openstack/nova stable/2024.2: Fix pci_tracker.save to delete all removed devs https://review.opendev.org/c/openstack/nova/+/959311
17:50:55 opendevreview sean mooney proposed openstack/nova master: hypervisors: Optimize uptime retrieval for better performance https://review.opendev.org/c/openstack/nova/+/959604
17:52:00 sean-k-mooney dan ^ implement the defaulting in the resouce tracker you wnated. i can update it more if needed
17:53:45 sean-k-mooney well technially not in the resouce trracker

Earlier   Later