Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2022-04-01
22:14:57 melwitt gmann: thanks. guh, weird that it's acting different now. I'll debug it
22:18:50 gmann yeah, I tried to debug but could not find much detail on why
23:41:32 opendevreview melanie witt proposed openstack/placement master: Make perfload jobs fail if write allocation fails https://review.opendev.org/c/openstack/placement/+/828438
#openstack-nova - 2022-04-02
00:18:39 opendevreview David Hill proposed openstack/nova master: Users can cheat --limit by setting -1 https://review.opendev.org/c/openstack/nova/+/836219
17:30:23 elodilles melwitt: so far i only saw that lc is broken on xena as it seems py36 is not present anymore in ubuntu-focal nodes (newer and older branches are not impacted beause we pin python3.6 in lc job in xena. which means actually that it should be broken on wallaby and victoria as well since we are not testing lc there with py36) but i haven't found yet where was removed python3 from ubuntu-focal node
17:30:29 elodilles build
#openstack-nova - 2022-04-03
05:58:50 melwitt elodilles: sorry I didn't link you earlier but it was from this https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/834179
05:59:22 melwitt here's the discussion where we're talking about dropping l-c job from such branches https://meetings.opendev.org/irclogs/%23openstack-infra/%23openstack-infra.2022-04-01.log.html#t2022-04-01T21:58:59
06:57:15 elodilles melwitt: thx! i think that with the new pip resolver and considering we already fixed all issues that were hidden before due to failing behavior of the resolver - maintaining of the l-c job is not that big of a problem. (we will have more issues with setuptools, virtualenv, tox, pip, etc. other things which mostly affect all jobs). anyway, i also think that the l-c is mostly necessary (for
06:57:21 elodilles packagers) at official release time, and probably not that relevant under stable (anyone can search back what was the l-c when it was dropped at branch cut)
06:57:45 elodilles melwitt: so in short, i'm OK with dropping the l-c on all stable branches, as some of the projects are already doing o:)
07:00:15 melwitt elodilles: heh, cool. well, I'm fine with either way. if you know how to fix it and would prefer it, definitely feel free :)
07:08:28 elodilles melwitt: i'll have a quick look on Monday, but to tell you the truth i don't want to go against the tide :D so we can simply drop them from stable branches and we can forget about it o:)
07:21:52 melwitt elodilles: tbh there isn't really a tide in nova, we just want it to not block the stable gates :P
15:16:18 opendevreview Takashi Kajinami proposed openstack/nova master: Fix wrong attribute to find remote address https://review.opendev.org/c/openstack/nova/+/836248
15:22:46 opendevreview Takashi Kajinami proposed openstack/nova master: Fix wrong attribute to find remote address https://review.opendev.org/c/openstack/nova/+/836248
15:45:10 opendevreview Takashi Kajinami proposed openstack/nova master: Deprecate [DEFAULT] use_forwarded_for https://review.opendev.org/c/openstack/nova/+/836253
15:49:14 opendevreview Takashi Kajinami proposed openstack/nova master: Deprecate [DEFAULT] use_forwarded_for https://review.opendev.org/c/openstack/nova/+/836253
18:14:16 opendevreview Takashi Kajinami proposed openstack/nova master: Deprecate [api] use_forwarded_for https://review.opendev.org/c/openstack/nova/+/836253
#openstack-nova - 2022-04-04
12:38:45 yuval Hey how do I join the PTG I need to download the eventbrite app?
12:42:03 sean-k-mooney yuval: no
12:42:10 sean-k-mooney yuval: https://ptg.opendev.org/ptg.html
12:42:15 sean-k-mooney this is the ptg site
12:42:36 sean-k-mooney if you click on the project in the calandar it will open the zoom meeting for you
12:42:49 sean-k-mooney there is also the list of etherpads at the top
12:43:03 sean-k-mooney some will use jitsi meet instead of zoom
12:43:23 sean-k-mooney but you just need a web browzer to join
12:43:32 sean-k-mooney no need to install the client or download eventbrite
12:48:10 kashyap gibi[m]: Just to give a heads-up (for later): on this (old) CPU APIs patch, I've addressed some comments from Stephen, and I'm systematically combing through some of your older remarks - https://review.opendev.org/c/openstack/nova/+/762330/
12:48:34 kashyap gibi[m]: ... and adress what I can. I'll include a short changelog of summary when I post a revison, so you know what's addressed
12:48:45 yuval thanks
12:48:46 kashyap gibi[m]: Don't look at it yet, but just giving a heads-up for later :)
12:50:33 gibi kashyap: ack, thanks for the headsup
13:01:27 bauzas happy PTG for everyone
13:02:26 gibi bauzas: o/
13:08:07 bauzas I have nothing specific in the PTG agenda that I want to go, but ping me if you want me to be around
13:08:53 bauzas gmann: do you want me to be around on the QA meeting given you would discuss about Secure RBAC ?
13:18:21 gmann bauzas: not as such from nova perspective but you can join on what plan we will do
13:18:46 bauzas gmann: ok, b/c we have a TC/community session at 2pm UTC
13:21:42 gmann bauzas: yeah, there we need you. but TC one will be after QA RBAC
13:21:54 bauzas gmann: I'm around on QA
13:22:01 gmann cool
13:22:07 bauzas gmann: then I'll go to the TC meeting at 2pm
13:26:30 gmann bauzas: thanks.
13:35:46 kashyap gibi: I have a question on an old comment of yours here on line-795: https://review.opendev.org/c/openstack/nova/+/762330/20..23/nova/virt/libvirt/host.py
13:36:28 kashyap gibi: You're saying about capabilities being logged but then we modify. I just want to make sure you've noticed the two libvirt APIs in play here:
13:37:01 kashyap - getCapabilities will tell you what libvirt code can do in general ("yes, this libvirt is new enough for particular XML constructs")
13:37:12 kashyap - getDomainCapbilities will tell you what libvirt can do with the given qemu-kvm binary
13:38:58 kashyap Isn't the logging being done there is of getCapabilties(), but the modification on line 809/810 (PS20) is being done of _get_domain_capabilities() - the second API?
13:49:57 gibi kashyap: I have to look at it again
13:50:04 gibi give me some time :)
13:50:12 kashyap Yeah, new context. If you're busy w/ PTG stuff, let it wait
13:50:37 kashyap Right, I'm posting a new rev (not touching this part yet). Also closely read the note from chengsheng, I think he's confusing. (Or I'm doing it)
13:51:51 gibi yeah, it was hard to follow
14:00:08 gibi kashyap: yeah, I think that is what I thought when I wrote the comment.
14:00:20 kashyap gibi: You're right
14:00:33 gibi kashyap: looking at master we already do that in case of arm
14:00:47 gibi to add cpu features
14:00:50 kashyap gibi: Already do what, sorry? Can you give a code pointer
14:01:30 gibi https://github.com/openstack/nova/blob/b0851b0e9c82446aec2ea0317514766fbc53abc0/nova/virt/libvirt/host.py#L806
14:01:42 gibi we already modify self._caps after we logged it
14:02:15 kashyap Ahh, so it is just "following the existing behaviour" from Ari
14:02:18 kashyap s/Ari/Arm/
14:02:36 gibi it seems like so
14:02:44 gibi so feel free to ignore that comment
14:03:08 kashyap Right; I'll leave it intact in this case.
14:03:16 gibi OK
14:03:39 kashyap I'm addressing your other comments. Thanks for the quick look
14:06:32 gibi thanks for picking that patch up
14:23:38 kashyap gibi: I might need some more help to get it to the finish line. This is getting more important as I see bugs in this area that can be fixed by these APIs
16:07:55 opendevreview Kashyap Chamarthy proposed openstack/nova master: Libvirt: Use improved guest CPU config APIs https://review.opendev.org/c/openstack/nova/+/762330
16:15:04 sean-k-mooney bauzas: have you finished organising the ptg adgenda by the way?
16:15:13 bauzas sean-k-mooney: yup
16:15:20 sean-k-mooney cool
16:16:01 sean-k-mooney we are starting at 14:00 utc ok
16:16:03 gibi bauzas: do we want to add a separate agenda point for the blind recheck discussion or we handle that as part of the nova's CI footprint topic?
16:16:19 sean-k-mooney neutron i think are starting at 13:00 so i will likely be intere room before we start
16:16:22 bauzas gibi: good question, I saw your addition
16:16:37 bauzas gibi: I'm ok with leaving it as one topic
16:16:50 bauzas gibi: I don't have anything more to say besides it
16:17:00 sean-k-mooney gibi: we used to have contibutor docs for recheck i dont think it will be contoverl to re add them to nova
16:17:15 gibi sean-k-mooney: that is a good action item
16:17:26 sean-k-mooney so i hope its more an FYI and then proceed to the gerrit doc review for main details
16:17:26 gibi bauzas: OK, lets keep it as is then
16:17:35 bauzas then let's move the CI point closer to the procedural section
16:17:38 bauzas gibi: ^
16:17:53 gibi bauzas: I have no problem with that
16:18:04 bauzas done
16:18:22 gibi sean-k-mooney: top of the doc change I think we should start pointing to that doc when we see blind rechecks (i.e. apply peer pressure)
16:18:26 gibi bauzas: cool
16:19:20 bauzas gibi: yup, I think we agreed on ideally having a zuul reply that would provide this link but at least we can comment using this link
16:19:35 bauzas but let's not start the conversation by now
16:19:39 gibi yepp, I would you the link in my comment :)
16:19:42 gibi ack
16:21:05 sean-k-mooney gibi: we currently point to the shared contibutor doc but its kind of burried
16:21:22 sean-k-mooney so the shared one already tells peopel to not do a blind recheck
16:21:34 sean-k-mooney but again its not trivial to find that
16:23:22 bauzas ideally this should be a contributor doc, not a project doc
16:23:24 sean-k-mooney so here https://docs.openstack.org/nova/latest/contributor/code-review.html we have https://docs.openstack.org/infra/manual/developers.html#peer-review as the first line
16:23:31 sean-k-mooney bauzas: it is

Earlier   Later