| *** soniya is now known as soniya29 | 04:00 | |
| ianw | sigh, apart from spending *waaaay* too long debugging a stray :1 that screwed up all ipv6 networking, i'm back :/ | 04:21 |
|---|---|---|
| *** pojadhav- is now known as pojadhav | 04:40 | |
| *** ysandeep|out is now known as ysandeep | 05:12 | |
| *** bhagyashris_ is now known as bhagyashris | 05:52 | |
| hrw | morning | 06:16 |
| hrw | TB dock... I use lenovo t490s with company provided lenovo tb dock. works fine | 06:18 |
| hrw | ianw: where can I check did centos-stream-9 images got built for aarch64? waiting for centos-9-stream-arm64 nodes ;d | 06:19 |
| ianw | hrw: so far the dock is fine, but my next problem now I have ipv6 back is that my monitor doesn't have a HDMI input, so I need to go find a displayport cable somewhere :) | 06:39 |
| hrw | ;) | 06:39 |
| hrw | hdmi as legacy ;D | 06:40 |
| ianw | although the number of extra cables going in and out of this thing over just the simple wedge the t460s used to sit in makes me wonder if we're moving forward or backwards | 06:40 |
| ianw | anyway, images ... :) | 06:40 |
| ianw | all arm64 images build on nb03 | 06:40 |
| ianw | hence all logs @ https://nb03.opendev.org/ | 06:41 |
| ianw | https://nb03.opendev.org/centos-9-stream-arm64-0000000001.log looks good | 06:41 |
| hrw | thanks | 06:42 |
| ianw | are there more outstanding changes? sorry i've been a bit distracted today | 06:43 |
| hrw | https://review.opendev.org/q/topic:add-centos-stream-9-stuff from yesterday | 06:43 |
| hrw | cs9 arm64 nodes, added epel to cs9 | 06:44 |
| hrw | waiting for nodes to happen and then python wheel mirror can start building | 06:44 |
| ianw | ++ | 06:44 |
| ianw | i don't think i've done x86 centos-9 wheels yet. if it can wait till tomorrow i can get both going? | 06:45 |
| hrw | ianw: so - are nodes created on first demand (so I need to recheck to make them) or they are first made and then jobs can run? | 06:45 |
| hrw | ianw: cs9 lacks wheels on any arch. AFS was needed (or is still needed?) and then jobs merged and started | 06:46 |
| ianw | that depends on min-ready settings in nodepool, but yeah either way a recheck gets nodes going | 06:46 |
| hrw | started recheck in https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/836829 to see | 06:47 |
| hrw | centos-9-stream-arm64 node started | 06:50 |
| hrw | x86-64 failed on enabling epel to isntall 'parallel'. lets see aarch64 | 06:51 |
| ianw | ohhh, yes the driver script does use that | 06:53 |
| hrw | images have epel component enabled but behave like it is not there | 06:54 |
| ianw | hrm that install expicitly enables epel i think | 06:57 |
| hrw | yep | 06:58 |
| ianw | https://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/playbooks/wheel-cache/pre.yaml#L6 | 06:59 |
| hrw | && "${DIB_RELEASE%-stream}" -lt '9' | 06:59 |
| hrw | ianw: https://paste.centos.org/view/ce542afb is probably what we want | 07:00 |
| opendevreview | Marcin Juszkiewicz proposed openstack/diskimage-builder master: CentOS Stream 9 has EPEL now https://review.opendev.org/c/openstack/diskimage-builder/+/836905 | 07:01 |
| ianw | yeah i agree. that's going to require a release and nodepool bump unfortunately | 07:02 |
| *** jpena|off is now known as jpena | 07:02 | |
| hrw | no problem | 07:02 |
| hrw | kolla cs9 builds work without cache too | 07:03 |
| *** pojadhav is now known as pojadhav|lunch | 07:35 | |
| *** pojadhav|lunch is now known as pojadhav | 08:08 | |
| *** ysandeep is now known as ysandeep|lunch | 09:07 | |
| *** rlandy|out is now known as rlandy | 10:24 | |
| *** ysandeep|lunch is now known as ysandeep | 10:31 | |
| opendevreview | Merged openstack/diskimage-builder master: CentOS Stream 9 has EPEL now https://review.opendev.org/c/openstack/diskimage-builder/+/836905 | 10:45 |
| opendevreview | Merged openstack/project-config master: Add ansible-collection-kolla to projects.yaml https://review.opendev.org/c/openstack/project-config/+/836763 | 10:57 |
| *** dviroel|afk is now known as dviroel\ | 11:13 | |
| *** dviroel\ is now known as dviroel | 11:13 | |
| *** ysandeep is now known as ysandeep|afk | 12:17 | |
| *** timburke_ is now known as timburke | 12:57 | |
| *** ysandeep|afk is now known as ysandeep | 13:06 | |
| *** artom__ is now known as artom | 14:12 | |
| *** dviroel is now known as dviroel|ruck | 14:14 | |
| *** dviroel|ruck is now known as dviroel|ruck|lunch | 15:28 | |
| *** ysandeep is now known as ysandeep|out | 15:34 | |
| *** jpena is now known as jpena|off | 16:05 | |
| *** dviroel|ruck|lunch is now known as dviroel|ruck | 16:13 | |
| clarkb | looks like tripleo centos 7 jobs are retry failuring on some qemu package install ? | 16:55 |
| TheSov | so it looks like after testing with a firend | 17:43 |
| TheSov | no new ssh keys work in opendev | 17:43 |
| TheSov | i would ask that someone else try it | 17:43 |
| TheSov | old keys work, new does not | 17:44 |
| clarkb | I think most of us lack windows to test wsl with. We do test gerrit regularly with generated keys in our testing | 17:44 |
| clarkb | TheSov: are they -sk keys? | 17:45 |
| TheSov | -sk? | 17:45 |
| TheSov | they are normal ssh-keygen keys | 17:45 |
| clarkb | they are new keytypes that dintegrate with security keys like yubikeys | 17:45 |
| clarkb | gerrit doesn't support them | 17:45 |
| TheSov | is it possible for you to try generating a new key, seperate from your old one to test? | 17:46 |
| clarkb | on linux I can | 17:46 |
| fungi | TheSov: what options are you passing to ssh-keygen? and what do you answer to its prompts? | 17:48 |
| TheSov | none, i press enter for all | 17:49 |
| TheSov | no passphrase | 17:49 |
| TheSov | clean ubuntu install | 17:49 |
| fungi | that's not especially helpful without knowing what the defaults are for the prompts in your version of ssh-keygen | 17:49 |
| clarkb | that will likely generate an rsa key which may no work if openssh is newer | 17:50 |
| fungi | yes, if you are creating an rsa key (which is probably default) then that's likely the problem | 17:50 |
| clarkb | I just generated a key using `ssh-keygen -t ed25519 -f ./test-new-key` and it works | 17:50 |
| TheSov | so no rsa keys are supported? | 17:50 |
| clarkb | TheSov: it depends on your client software | 17:50 |
| fungi | ssh-rsa exchages are not supported by newer versions of the openssh client | 17:51 |
| TheSov | let me try your way | 17:51 |
| fungi | and it won't automatically try newer hashes with rsa keys, so it doesn't realize that gerrit supports them | 17:51 |
| clarkb | and current gerrit only knows how to negotiate that. When version 3.6 arrive sit shoul dsupport rsa-sha2 variants | 17:51 |
| clarkb | also remember that gerrit listens on port 29418 not the default of 22 | 17:52 |
| fungi | yes, using an elliptic curve key instead, like ed25519, seems to be a viable workaround in those cases | 17:52 |
| clarkb | so when you test you have to do something like ssh -i path-to-key -p 29418 username@review.opendev.org gerrit ls-projects | 17:52 |
| fungi | if you drop the "gerrit ls-projects" you should get a welcome banner like "**** Welcome to Gerrit Code Review ****" | 17:54 |
| TheSov | made the same type of key | 17:55 |
| TheSov | ssh -i ~/.ssh/id_ed25519 *MYUSER*@review.opendev.org | 17:55 |
| TheSov | fails wrong key | 17:55 |
| clarkb | you need the -p 29418 | 17:56 |
| TheSov | oh i see | 17:56 |
| TheSov | ok so i need to put that same port number in git-review? | 17:56 |
| fungi | also make sure you upload your new .ssh/id_ed25519.pub to your settings in the gerrit webui | 17:56 |
| TheSov | already did that | 17:56 |
| clarkb | no git review should know that gerrit is on port 29418 | 17:56 |
| TheSov | it worked on the new port | 17:56 |
| fungi | TheSov: git-review looks at a .gitreview file in the repository which already specifies that port number | 17:57 |
| TheSov | how do i specify a new key to git-review, would you know by any chance? | 17:57 |
| TheSov | I just want to thank both of you, you are both amazing! | 17:58 |
| fungi | you can edit your ~/.ssh/config and add something like: | 17:58 |
| fungi | Host review.opendev.org | 17:58 |
| clarkb | ya git just uses your normal ssh configs | 17:58 |
| fungi | IdentityFile ~/.ssh/id_ed25519 | 17:58 |
| clarkb | I load a couple of keys into my agent and it tries them | 17:58 |
| fungi | but yes, ssh will try all the keys it knows about in your ~/.ssh/ directory | 17:59 |
| fungi | and git uses ssh, and git-review uses git (and ssh) | 18:00 |
| clarkb | as noted before we expect gerrit version 3.6 to fix the rsa problems | 18:01 |
| clarkb | but that release hasn't happened yet and we have to get onto 3.5 first | 18:01 |
| clarkb | looks lik we can make a dib 3.20.2 release now. The only two changes since 3.20.1 are ubuntu imgae downloads with https and the epel 9 fixup | 18:16 |
| *** rlandy is now known as rlandy|mtg | 19:53 | |
| *** dviroel|ruck is now known as dviroel|afk | 20:39 | |
| ianw | clarkb: cool, yep did you do a release? | 20:41 |
| fungi | not yet it doesn't appear | 20:45 |
| Clark[m] | Sorry no | 20:47 |
| Clark[m] | I got distracted by food | 20:48 |
| ianw | no probs, i can push on that | 20:48 |
| ianw | pushed tag, i'll look at moving that through and the cs9 wheel builds today | 20:49 |
| *** rlandy|mtg is now known as rlandy | 21:17 | |
| clarkb | ianw: we need to push a nodepool image update to pick that up? | 21:38 |
| fungi | reqs bump yeah | 21:39 |
| clarkb | I'll push that | 21:39 |
| clarkb | remote: https://review.opendev.org/c/zuul/nodepool/+/837035 Update DIB to version 3.20.2 | 21:42 |
| ianw | ++ | 21:57 |
| hrw | ++ | 21:59 |
| opendevreview | Tim Burke proposed opendev/git-review master: Treat ^C as "no" at confirmation prompt https://review.opendev.org/c/opendev/git-review/+/837044 | 23:39 |
| ianw | why do we create ro replicas of afs replicas on the same server as the rw volume? i notice the extant centos8 volumes don't have it. i forget who set these up, it was probably me | 23:56 |
| ianw | i wonder if it means that /afs/openstack.org/mirror/wheel/centos8* have only been being served by the ord server? | 23:59 |
| ianw | (as opposed to /afs/.openstack/...) | 23:59 |
| ianw | we've got ourselves into a tangle with the naming conventions now | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!