Kubernetes ConfigMap volumes contain a .data directory, eg:
/ # ls -la /etc/prometheus
total 12
drwxrwxrwx 3 root root 4096 May 24 08:03 .
drwxr-xr-x 6 root root 4096 May 24 08:03 ..
drwxr-xr-x 2 root root 4096 May 24 08:03 ..2018_05_24_08_03_36.264564876
lrwxrwxrwx 1 root root 31 May 24 08:03 ..data -> ..2018_05_24_08_03_36.264564876
lrwxrwxrwx 1 root root 26 May 24 08:03 prometheus.tmpl.yml -> ..data/prometheus.tmpl.yml
When using a ConfigMap volume mount as reloader.rule-dir , reloader fails and sidecar crashes :
level=error ts=2018-05-24T11:47:54.472408145Z caller=main.go:147 msg="running command failed" err="build hash: read /etc/prometheus-rules/..data: is a directory"
Thanks! More testing needed. PRs welcome (:
Hi @bwplotka @clmssz,
Has anybody started working on this? If no I would like to.
Hi @szalai1, I think noone is working on it and I can confirm I see the log all the time.
If you want it would be great to solve this issue! So feel free to create PR :)
we recently wrote some code to handle this, in case someone is interested in transferring the approach: https://github.com/brancz/kube-rbac-proxy/pull/45
Great thanks for the info! I'll take a look at this if noone will pick it.
@FUSAKLA I didn't finish the work on this issue. So you can work on it.
FYI we had the same issue and the workaround is to set - --reloader.rule-dir=/etc/prometheus-rules/..data .
We've been running like that for a while now and it works flawlessly.
Great, will try it out. Thanks!
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
I believe this still exists as originally reported.
@brancz @FUSAKLA Test latest reloader pkg with following dir and can not reproduce:
[test]# ls -al
total 4
drwxr-xr-x 3 root root 45 1月 15 17:40 .
drwxr-xr-x 16 1000 1000 4096 1月 15 17:18 ..
drwxr-xr-x 2 root root 6 1月 15 17:19 ..aaa
lrwxrwxrwx 1 root root 5 1月 15 17:19 ..data -> ..aaa
This issue/PR has been automatically marked as stale because it has not had recent activity. Please comment on status otherwise the issue will be closed in a week. Thank you for your contributions.