ceph - sepia - 2024-06-17

Timestamp (UTC)Message
2024-06-17T04:35:52.839Z
<Leonid Usov> 🤦 <https://github.com/PyCQA/pycodestyle/commit/3cedd4c74b5f467fc4c1cde5a8ac2503a54b60db>
2024-06-17T04:42:28.098Z
<Leonid Usov> <https://github.com/PyCQA/flake8/commit/65a38c42a7f1a05ff8d99b313160754fc9b7a0d8>

flake8 released 7.1.0 yesterday, referencing pycodestyle with the new rule
2024-06-17T04:47:27.888Z
<Leonid Usov> I don’t see how we limit this version of flake8:
```flake8: install_deps /home/jenkins-build/build/workspace/ceph-pull-requests/src/pybind/mgr> python -I -m pip install flake8```
this command is run by … flake8??
2024-06-17T04:55:59.085Z
<Leonid Usov> OK I figured it out, it can be constrained in `src/pybind/mgr/tox.ini`.
```diff --git a/src/pybind/mgr/tox.ini b/src/pybind/mgr/tox.ini
index 53a0a776626..2e918b896ea 100644
--- a/src/pybind/mgr/tox.ini
+++ b/src/pybind/mgr/tox.ini
@@ -159,7 +159,7 @@ commands =
 
 [testenv:flake8]
 deps =
-    flake8
+    flake8<7.1.0
 allowlist_externals = bash
 modules =
     alerts \```
Now the question is, should we fix the new warning, or should we pin the version of flake8?
Even if we fix those whitespaces, we should still probably freeze the flake version to avoid breakages like this one.
2024-06-17T04:57:37.194Z
<Leonid Usov> OK I figured it out, it can be constrained in `src/pybind/mgr/tox.ini`.
```diff --git a/src/pybind/mgr/tox.ini b/src/pybind/mgr/tox.ini
index 53a0a776626..2e918b896ea 100644
--- a/src/pybind/mgr/tox.ini
+++ b/src/pybind/mgr/tox.ini
@@ -159,7 +159,7 @@ commands =
 
 [testenv:flake8]
 deps =
-    flake8
+    flake8<7.1.0
 allowlist_externals = bash
 modules =
     alerts \```
_(this patch wasn’t tested!)_
Now the question is, should we fix the new warning, or should we pin the version of flake8?
Even if we fix those whitespaces, we should still probably freeze the flake version to avoid breakages like this one.
2024-06-17T05:17:09.634Z
<nehaojha> @Dan Mick what can we do about them?
2024-06-17T07:20:19.615Z
<Pedro Gonzalez Gomez> For now, this was addressed here: <https://github.com/ceph/ceph/pull/58062>
2024-06-17T11:18:10.888Z
<Adam Kraitman> I plan to migrate tracker to a newer instance in the cloud what should speed up upload and also could speed up download time so I guess we will see after the migration if we still see those errors from time to time
2024-06-17T12:11:56.354Z
<Guillaume Abrioux> @Adam Kraitman I see there are still some jenkins worker nodes running el8 😕
2024-06-17T12:12:22.200Z
<Guillaume Abrioux> <https://jenkins.ceph.com/computer/172.21.5.32+adami02/>
2024-06-17T12:13:54.272Z
<Guillaume Abrioux> irvingi01
irvingi02
irvingi03
adami01
adami02
2024-06-17T12:20:41.182Z
<Guillaume Abrioux> I suspect the label has not been updated
2024-06-17T13:05:50.403Z
<John Mulligan> This whitespace thing is certainly not idiomatic python.
2024-06-17T13:05:56.929Z
<John Mulligan> Why can't we fix the code?
2024-06-17T13:06:05.452Z
<Ronen Friedman> @Leonid Usov @Pedro Gonzalez Gomez - thank you!
2024-06-17T13:07:41.709Z
<Leonid Usov> > Why can’t we fix the code?
we can and that’s what the PR above does. However, I still think we need to lock the version of flake8 to avoid cases when all PRs fail to pass CI just because there was a library update
2024-06-17T13:08:31.934Z
<John Mulligan> Sorry. I didn't see the PR (long threads suck in Slack (IMO)).
2024-06-17T13:09:36.329Z
<John Mulligan> The downside to pinning a version is that someone needs to look at the pinned versions. We can not just pin a version forever - we need those fixes, improvement, and eventually those old versions may not be compatible with the base os, python version etc.
2024-06-17T13:10:59.043Z
<Adam Kraitman> I have those left I hope to finish soon <https://jenkins.ceph.com/label/centos8/>
2024-06-17T14:21:30.539Z
<gregsfortytwo> @Adam Kraitman as well
2024-06-17T14:21:58.758Z
<gregsfortytwo> I thought ceph-users was moderated for non-members, which is why it hasn’t seen this problem in a very long time — despite sepia@ (not moderated) getting hit hard recently
2024-06-17T14:22:18.831Z
<gregsfortytwo> did some spammers actually get approved accounts before they went to town, or did some configuration get changed inadvertently?
2024-06-17T16:26:51.677Z
<Laura Flores> Thanks! Looks great!
2024-06-17T16:50:43.008Z
<Adam Kraitman> Thanks !
2024-06-17T18:24:52.947Z
<alvaro.soto> 2cents, what you guys think about moving to google groups?
2024-06-17T18:26:11.508Z
<alvaro.soto> since it's getting hared and harder to get admins for community related apps, I do think it's a good idea to get out of self hosted apps
2024-06-17T18:27:11.767Z
<Dan Mick> I'm sort of stuck with [lists.ceph.io](http://lists.ceph.io) at the moment.  It's not logging sufficiently on the error to pinpoint anything, and it's fairly old code, so my first step was to try to upgrade, but I'll need to back up data/configs and that process is apparently somewhat manual.
2024-06-17T18:44:12.035Z
<alvaro.soto> is there even a path to migrate from v1 to v3?
2024-06-17T18:52:31.335Z
<Dan Mick> this is v3.3.1, but there is a later revision of the containers we're running it
2024-06-17T18:52:38.804Z
<Dan Mick> this is v3.3.1, but there is a later revision of the containers we're running in
2024-06-17T18:55:35.883Z
<alvaro.soto> "502 Bad Gateway" =/
2024-06-17T18:56:26.669Z
<alvaro.soto> what do you think about migrating to googlegroups and forget about all of the pain @Dan Mick?
2024-06-17T18:58:55.790Z
<Dan Mick> I don't know the factors that went into choosing mailman to begin with.  I'm chary about using Google; if they support what you want now, that doesn't promise anything about the future, and they can and do just shut services down with little warning
2024-06-17T19:04:11.110Z
<alvaro.soto> in the end they are just emails and the list of subscribers, so the main account @list.ceph.io will have always the backup, if google shutdown the groups service, then will be time to migrate to another self hosted product, don't you think ?
2024-06-17T19:33:10.968Z
<Dan Mick> I don't know what that means.  Transitioning requires effort, and the services aren't just emails and subscribers, there's also collected block lists and the archives.  I'm not even sure how those might transition to Groups.  And transitioning back is also a cost, of course.  I'm not saying it's a bad idea, just that it's not trivial and may involve some loss of service.
2024-06-17T19:36:35.701Z
<alvaro.soto> can you share all the version and apps behind [list.ceph.io](http://list.ceph.io)? in that way others (me) can take a look at options?
2024-06-17T19:37:18.039Z
<Dan Mick> It's Mailman v3.3.1
2024-06-17T19:37:37.251Z
<alvaro.soto> I mean, even google groups get spam, but I do believe we need to get out of selfhosted apps if we can
2024-06-17T19:41:39.431Z
<alvaro.soto> ty 😃
2024-06-17T20:52:32.452Z
<Laura Flores> @Adam Kraitman are you the right person to ask to check into the ceph-users list spam messages we've been getting lately?

Here is an example: <https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/thread/PKYMP46U7ADDODSC3CAJ4SBZVUKO5CKU/>

Some details from @David Orman (thanks!) that we discussed in the CLT call:
• SpamAssassin scoring seems unusually low for the types of emails coming through
• Currently version 3.4.6, which is relatively recent, but the scoring seems abnormally low for the content of the emails

2024-06-17T22:59:29.611Z
<alvaro.soto> has renamed the channel from "sepia" to "ceph-sepia"
2024-06-17T23:04:00.434Z
<khyr0n> (sorry about the noise, doing some reconfig)
2024-06-17T23:41:45.791Z
<Laura Flores> Hey @Dan Mick, we are close to being ready for the ARM CI check to be made required. (Still at least one outstanding bug to address first). In preparation for when we're ready, do you know how the check can be made required?
cc @Rongqi Sun
2024-06-17T23:42:24.562Z
<Dan Mick> no more specifically than "poke around Github"
2024-06-17T23:42:53.139Z
<Dan Mick> I would wait until it's passing constantly for a week or so.  Is it?
2024-06-17T23:43:31.387Z
<Laura Flores> Not yet, but we're on the way to that. Just asking in advance
2024-06-17T23:45:56.345Z
<Dan Mick> [https://docs.github.com/en/repositories/configuring-branches-and-merges-in-your-repository/managing-protected-branches/about-protected-branches#[…]g](https://docs.github.com/en/repositories/configuring-branches-and-merges-in-your-repository/managing-protected-branches/about-protected-branches#require-status-checks-before-merging)

Any issue? please create an issue here and use the infra label.