| -@gerrit:opendev.org- Michael Kelly proposed: | 06:09 | |
| - [zuul/zuul-jobs] 871679: roles: Add ensure-git-lfs https://review.opendev.org/c/zuul/zuul-jobs/+/871679 | ||
| - [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | ||
| @picog:matrix.org | Hi Folks. I've been upgrading a legacy CI system based on software-factory. | 09:32 |
|---|---|---|
| I upgrade from Software factory 3.4 to 3.5. As I understand it, the runc container now goes away. | ||
| After the upgrade, I see tox-linter jobs immediately fail with node failure. | ||
| Could someone perhaps help me understand where I should be looking to fix this. | ||
| Where does zuul figure out which provider to use when requesting a node from nodepool: | ||
| This is the log I see: | ||
| ``` | ||
| INFO zuul.nodepool: [e: cf2f7201b176428ebe701e565ef659cd] Submitted node request <NodeRequest 200-0000224653 <NodeSet [<Node None ('managed-runc-provider',):runc-centos>]>>``` | ||
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 837908: Generate a token silently if possible https://review.opendev.org/c/zuul/zuul-client/+/837908 | 09:35 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 837908: Generate a token silently if possible https://review.opendev.org/c/zuul/zuul-client/+/837908 | 10:07 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 842361: zuul-web: harmonize behavior for boolean parameters https://review.opendev.org/c/zuul/zuul/+/842361 | 10:17 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 794553: builds: fix API queries for boolean parameters, make tenant optional https://review.opendev.org/c/zuul/zuul-client/+/794553 | 10:20 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 819118: Support for "basic" authentication https://review.opendev.org/c/zuul/zuul-client/+/819118 | 10:22 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 819118: Support for "basic" authentication https://review.opendev.org/c/zuul/zuul-client/+/819118 | 10:27 | |
| -@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 871259: Filter input validation for "Change" https://review.opendev.org/c/zuul/zuul/+/871259 | 11:08 | |
| -@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 871259: Filter input validation for "Change" https://review.opendev.org/c/zuul/zuul/+/871259 | 11:13 | |
| @fungicide:matrix.org | > <@picog:matrix.org> Hi Folks. I've been upgrading a legacy CI system based on software-factory. | 13:04 |
| > I upgrade from Software factory 3.4 to 3.5. As I understand it, the runc container now goes away. | ||
| > After the upgrade, I see tox-linter jobs immediately fail with node failure. | ||
| > Could someone perhaps help me understand where I should be looking to fix this. | ||
| > Where does zuul figure out which provider to use when requesting a node from nodepool: | ||
| > | ||
| > This is the log I see: | ||
| > ``` | ||
| > INFO zuul.nodepool: [e: cf2f7201b176428ebe701e565ef659cd] Submitted node request <NodeRequest 200-0000224653 <NodeSet [<Node None ('managed-runc-provider',):runc-centos>]>>``` | ||
| without knowing what zuul and nodepool versions were in sf 3.4 and 3.5 it's hard to say what the upgrade implications might be. according to https://www.softwarefactory-project.io/ they have their own matrix channel and a mailing list you can ask for help on, that might be a good place to start | ||
| @picog:matrix.org | > <@fungicide:matrix.org> without knowing what zuul and nodepool versions were in sf 3.4 and 3.5 it's hard to say what the upgrade implications might be. according to https://www.softwarefactory-project.io/ they have their own matrix channel and a mailing list you can ask for help on, that might be a good place to start | 13:18 |
| Thanks. Zuul is now on 3.19 and nodepool 3.13. | ||
| I did ask on the software factory channel before asking here. | ||
| I found the zuul-jobs repo at least which contains the definitions of the linter jobs, I'll dig through that | ||
| @fungicide:matrix.org | Pico: they'll probably be in a much better position to assist. those versions of zuul and nodepool are several years old, so while i can point you to the release notes, i'm not sure i remember what the upgrade entailed that long ago | 13:21 |
| @fungicide:matrix.org | https://zuul-ci.org/docs/zuul/latest/releasenotes.html#relnotes-3-19-1 https://zuul-ci.org/docs/nodepool/latest/releasenotes.html#relnotes-3-13-1 | 13:23 |
| @picog:matrix.org | > <@fungicide:matrix.org> Pico: they'll probably be in a much better position to assist. those versions of zuul and nodepool are several years old, so while i can point you to the release notes, i'm not sure i remember what the upgrade entailed that long ago | 13:24 |
| Yeah, I've trying to incrementally upgrade I don't want to risk an upgrade from a broken system. | ||
| @picog:matrix.org | > <@fungicide:matrix.org> Pico: they'll probably be in a much better position to assist. those versions of zuul and nodepool are several years old, so while i can point you to the release notes, i'm not sure i remember what the upgrade entailed that long ago | 13:24 |
| * Yeah, I'm trying to incrementally upgrade I don't want to risk an upgrade from a broken system. | ||
| @fungicide:matrix.org | if it helps, we do try to record breaking changes and upgrade steps in the release notes when there are any you need to be aware of | 13:26 |
| @fungicide:matrix.org | and usually any significant breaking changes come along with a major version increment (for example when you get to zuul 4.x you'll need to make sure you have a relational database available and configured) | 13:27 |
| -@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 871259: Filter input validation for "Change" https://review.opendev.org/c/zuul/zuul/+/871259 | 13:35 | |
| @mhuin:matrix.org | Hello, I'm dusting up some old open changes for zuul-client, these two are adding auth features and got +2'ed a while ago: | 13:55 |
| @mhuin:matrix.org | 1 - https://review.opendev.org/c/zuul/zuul-client/+/819118 will allow a user to authenticate with username/password rather than a raw JWT | 13:55 |
| @mhuin:matrix.org | 2 - https://review.opendev.org/c/zuul/zuul-client/+/837908 will generate a token automatically if a zuul.conf file with a H256 authenticator is available. The use case is: using zuul-client on the server where zuul is running | 13:56 |
| @mhuin:matrix.org | They're conflicting so if you have a preference for one being merged before the other let me know so I can do the needed rebase | 13:57 |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul-client] 752909: Add buildsets, buildset-info to subcommands https://review.opendev.org/c/zuul/zuul-client/+/752909 | 14:47 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 758985: zuul-client: Test buildsets, buildset-info subcommands https://review.opendev.org/c/zuul/zuul/+/758985 | 14:55 | |
| @clarkb:matrix.org | corvus: ianw executors updated in opendev so I have rechecked https://review.opendev.org/c/zuul/nodepool/+/871409/ | 16:15 |
| -@gerrit:opendev.org- Michael Kelly proposed: [zuul/zuul] 869786: model: Eliminate semaphore_handler parameter on findJobsTo*() https://review.opendev.org/c/zuul/zuul/+/869786 | 17:08 | |
| -@gerrit:opendev.org- Michael Kelly proposed: | 17:10 | |
| - [zuul/zuul-jobs] 871679: roles: Add ensure-git-lfs https://review.opendev.org/c/zuul/zuul-jobs/+/871679 | ||
| - [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | ||
| @michael_kelly_anet:matrix.org | corvus: I added some roles to cover the LFS stuff we had chatted about last week as well as one to cover submodules. | 17:11 |
| @michael_kelly_anet:matrix.org | https://review.opendev.org/q/project:zuul/zuul-jobs+status:open+owner:mkelly | 17:11 |
| -@gerrit:opendev.org- Michael Kelly proposed: | 17:56 | |
| - [zuul/zuul-jobs] 871679: roles: Add ensure-git-lfs https://review.opendev.org/c/zuul/zuul-jobs/+/871679 | ||
| - [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | ||
| @clarkb:matrix.org | One of OpenDev's users `check experimental`'d an old patchset (patchset 14 of this change https://review.opendev.org/c/openstack/nova/+/863920 when patchset 15 was latest). Zuul appears to have run the jobs against patchset 14: https://zuul.opendev.org/t/openstack/builds?pipeline=experimental&change=863920%2C14&skip=0 but as you can see on the change it never reported back. | 19:00 |
| @clarkb:matrix.org | I tracked it down to Gerrit returning 404 responses when posting the review. This is where Zuul is posting the review `DEBUG zuul.GerritConnection: POST: https://review.opendev.org/a/changes/openstack%2Fnova~master~I595b27a57516cffe1f172cf2fb736e1b11373a1d/revisions/None/review` Note the `None` in the revision field. | 19:02 |
| @clarkb:matrix.org | And that url is constructed here: https://opendev.org/zuul/zuul/src/branch/master/zuul/driver/gerrit/gerritconnection.py#L1388-L1390 looks like change.commit is `None` for some reason. | 19:04 |
| @clarkb:matrix.org | I've got to run, but wanted to call this out. Maybe there is already a fix for this or someone understands why this would ahppen? | 19:04 |
| -@gerrit:opendev.org- Michael Kelly proposed: | 19:05 | |
| - [zuul/zuul-jobs] 871679: roles: Add ensure-git-lfs https://review.opendev.org/c/zuul/zuul-jobs/+/871679 | ||
| - [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | ||
| @clarkb:matrix.org | hrm Refs/Branches/Changes in model.py don't have a commit attribute | 19:09 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 19:09 | |
| - [zuul/nodepool] 871776: Fix openshift doc attribute hierarchy https://review.opendev.org/c/zuul/nodepool/+/871776 | ||
| - [zuul/nodepool] 871777: Add support for privileged containers https://review.opendev.org/c/zuul/nodepool/+/871777 | ||
| @clarkb:matrix.org | so now I'm really confused where that comes from (BuildSet does though_) | 19:09 |
| @clarkb:matrix.org | but I need to run errands back later. | 19:09 |
| -@gerrit:opendev.org- Michael Kelly proposed: | 19:26 | |
| - [zuul/zuul-jobs] 871679: roles: Add ensure-git-lfs https://review.opendev.org/c/zuul/zuul-jobs/+/871679 | ||
| - [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | ||
| @clarkb:matrix.org | GerritChange which inherits from Change adds that attribute | 20:30 |
| @clarkb:matrix.org | Knowing that I now realize something interesting. The ssh and http code paths have different behaviors. Compare https://opendev.org/zuul/zuul/src/branch/master/zuul/driver/gerrit/gerritmodel.py#L91-L94 to https://opendev.org/zuul/zuul/src/branch/master/zuul/driver/gerrit/gerritmodel.py#L151-L156 | 20:31 |
| @clarkb:matrix.org | I think that second link explains why the commit is None. It is because self.patchset != the current revision. But in the ssh path we would handle that | 20:31 |
| @clarkb:matrix.org | corvus: ^ not sure what the best way to reconcile those behavior differences is. I feel like the most correct thing is to update http to act like ssh? | 20:32 |
| @clarkb:matrix.org | aha when we query via ssh we set --patch-sets to get all patchests back but when we query by http we only set the flag for the current revision | 20:36 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 871783: WIP consistent patchset behavior for gerrit ssh and http https://review.opendev.org/c/zuul/zuul/+/871783 | 20:59 | |
| @clarkb:matrix.org | That is very much not correct or complete, but I think it should have enough hints in it to help when picking this up again later as I don't have time for it right now | 20:59 |
| -@gerrit:opendev.org- Michael Kelly proposed: [zuul/zuul-jobs] 871680: roles: Add git-lfs-init https://review.opendev.org/c/zuul/zuul-jobs/+/871680 | 21:48 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 22:59 | |
| - [zuul/nodepool] 871199: Add API timing debug statements to openstack driver https://review.opendev.org/c/zuul/nodepool/+/871199 | ||
| - [zuul/nodepool] 871803: Speed up openstack server listing https://review.opendev.org/c/zuul/nodepool/+/871803 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 23:47 | |
| - [zuul/nodepool] 871803: Speed up openstack server listing https://review.opendev.org/c/zuul/nodepool/+/871803 | ||
| - [zuul/nodepool] 871199: Add API timing debug statements to openstack driver https://review.opendev.org/c/zuul/nodepool/+/871199 | ||
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!