| clarkb | fungi: I've approved the first change in that stack which should address the base jobs issue. | 00:23 |
|---|---|---|
| fungi | thanks! | 00:28 |
| ianw | clarkb: [2023-02-14T00:22:24.158Z] [sshd-SshDaemon[11a8735d](port=22)-nio2-thread-2] DEBUG com.google.gerrit.sshd.ChannelIdTrackingUnknownChannelReferenceHandler : channelInitialized(ChannelSession[id=0, recipient=-1]-ServerSessionImpl[admin@/127.0.0.1:35080]) updated last tracked channel ID null => 0 | 00:29 |
| ianw | amazing what happens when you actually register the handler :) | 00:29 |
| clarkb | that was it then | 00:29 |
| fungi | d'oh! | 00:30 |
| ianw | this also probably "fixes" the problem, in so much as the original "fixed" anything. because the channel tracker never initialized anything, it's logic of ignoring the channel error if the channel had ever been opened never triggered | 00:31 |
| ianw | i'm going to talk a quick walk, but will reorganise all the changes and reply to the original bug | 00:33 |
| clarkb | ianw: I think I understand what you are saying but having a hard time parsing that :) | 00:33 |
| clarkb | the channel isn't being found because they didn't fully register the tracker whcih elads to the ssh connection issues? | 00:33 |
| clarkb | but only if it needs to lookup the channel tracker? | 00:34 |
| clarkb | enjoy your walk I can read the bug update :) | 00:34 |
| ianw | yeah, basically "lastTracked" in "if ((lastTracked != null) && (channelId <= lastTracked.intValue()))" is never != null | 00:36 |
| opendevreview | Merged opendev/system-config master: Farewell limestone https://review.opendev.org/c/opendev/system-config/+/873428 | 02:23 |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] gerrit: increase ssh channel debugging https://review.opendev.org/c/opendev/system-config/+/873214 | 02:41 |
| *** yadnesh|away is now known as yadnesh | 04:28 | |
| *** jpena|off is now known as jpena | 08:25 | |
| *** gibi_ is now known as gibi | 09:00 | |
| *** jpena is now known as jpena|off | 09:39 | |
| *** jpena|off is now known as jpena | 10:06 | |
| *** dviroel|out is now known as dviroel | 10:58 | |
| *** dasm|off is now known as dasm | 13:10 | |
| *** yadnesh is now known as yadnesh|away | 13:31 | |
| *** \join_subline is now known as \join_weakmayors | 14:36 | |
| clarkb | no new cert complaints! | 16:11 |
| fungi | excellent! | 16:15 |
| fungi | infra-root: config-core: any additional opinions on caching a new version of cirros on our nodes? https://review.opendev.org/873244 | 17:19 |
| clarkb | let me double check urls and I can approve it | 17:20 |
| clarkb | we may be able to remove some old cirros images too | 17:20 |
| fungi | see my comments on the review, yeah | 17:20 |
| fungi | also 873429 and 873430 are more cleanup changes for clouds we no longer use for nodepool | 17:21 |
| clarkb | fungi: ya I think even if 0.3.4 is still used we probably download it once a month and can just let the jobs do that at runtime | 17:21 |
| fungi | that was my thinking too. if opensearch isn't showing a version as used in its retention period, then it's probably not necessary to cache | 17:22 |
| fungi | which means we could drop everything except 0.5.2 while adding 0.6.1 | 17:22 |
| fungi | so 0.3.4, 0.3.5, 0.4.0 | 17:23 |
| fungi | and 0.5.1 | 17:23 |
| fungi | i'll propose that as a followup change | 17:24 |
| *** jpena is now known as jpena|off | 17:24 | |
| clarkb | fungi: the inap/iweb/packethost cleanups have been approved | 17:26 |
| fungi | thanks! | 17:26 |
| clarkb | I shold actually try to dedicate a good chunk of today to reviews. I feel like I'm falling way behind on those | 17:26 |
| opendevreview | Jeremy Stanley proposed openstack/project-config master: Stop caching infrequently-used CirrOS images https://review.opendev.org/c/openstack/project-config/+/873735 | 17:28 |
| opendevreview | Merged openstack/project-config master: Cache Cirros 0.6.1 images https://review.opendev.org/c/openstack/project-config/+/873244 | 17:34 |
| opendevreview | Merged opendev/system-config master: Finish cleaning up packethost references https://review.opendev.org/c/opendev/system-config/+/873429 | 18:07 |
| opendevreview | Merged opendev/system-config master: Final cleanup of internap/inap/iweb references https://review.opendev.org/c/opendev/system-config/+/873430 | 18:07 |
| opendevreview | Merged opendev/system-config master: mailman: set web auto field size to "AutoField" https://review.opendev.org/c/opendev/system-config/+/873337 | 19:45 |
| clarkb | ianw: I've approved the docker signed-by key change | 20:01 |
| ianw | thanks i'll keep an eye on deployment | 20:01 |
| clarkb | left a note in there for some quick version support I dug up. Seems like as far back as xenial may do it even | 20:01 |
| fungi | keep in mind there was a point at which binary pgp files were supported there but asc encoded file support came later | 20:03 |
| opendevreview | Merged opendev/system-config master: install-docker: switch from deprecated apt-key https://review.opendev.org/c/opendev/system-config/+/872808 | 21:03 |
| clarkb | ianw: looks like codesearch deploy failed on ^ | 21:34 |
| clarkb | though apparently I never loaded ssh keys so I can't pull logs up quickly | 21:34 |
| clarkb | codesearch itself is still running though | 21:34 |
| clarkb | ianw: Failed to update apt cache: unknown reason | 21:36 |
| fungi | 104.130.172.34 is the new held lists99 | 21:37 |
| clarkb | I think I see the issue. Trying to confirm on codesearch then will push a fix | 21:38 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Fix docker.gpg.asc path in sources.list entry https://review.opendev.org/c/opendev/system-config/+/873754 | 21:41 |
| clarkb | ianw: fungi ^ I think that corrects it | 21:41 |
| clarkb | I did that by hand on codesearch and `sudo apt-get update` was happy afterwards | 21:41 |
| fungi | usually the .gpg extension means binary format and .asc is ascii-encoded | 22:01 |
| clarkb | makes sense. In this case it is ascii encoded so .asc is correct we just have to point at it properly | 22:01 |
| fungi | i would have expected it to be either docker.gpg or docker.asc, in retrospect, but fixing it to be consistent with the configuration so it works is what matters for now yes | 22:03 |
| clarkb | oh you mean .gpg and .asc instead of .gpg and .gpg.asc | 22:07 |
| fungi | yes, .asc is generally interpreted as ascii-armored gpg data in that context | 22:15 |
| ianw | clarkb: thanks! ... how did all the tests pass with that? | 23:23 |
| ianw | WARNING: The following packages cannot be authenticated! | 23:25 |
| ianw | containerd.io docker-ce-cli docker-ce | 23:25 |
| ianw | Authentication warning overridden. | 23:25 |
| ianw | https://b30211aa4f809fc4a91b-baf4f807d40559415da582760ebf9456.ssl.cf2.rackcdn.com/872808/1/gate/system-config-run-codesearch/c35301e/bridge99.opendev.org/ara-report/results/313.html | 23:25 |
| ianw | "allow_unauthenticated": false, is disabled | 23:30 |
| ianw | so why was it allowed?! | 23:30 |
| clarkb | ianw: oh because we don't sign our mirrors either and we disable that | 23:32 |
| clarkb | via /etc/apt/preferences.d or soething | 23:32 |
| ianw | hrm, yeah | 23:34 |
| ianw | APT::Get::AllowUnauthenticated | 23:34 |
| ianw | i think now we can specify that in the sources line | 23:36 |
| ianw | 23.04 is also going to switch to deb822 format, which is going to break ~ everything | 23:48 |
| clarkb | thats for the sources.list files? | 23:49 |
| ianw | sources.list will go with individual files in /etc/apt/sources.list.d, in the more verbose format, aiui | 23:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!