Project

General

Profile

Bug #58661

Deploying SSL radosgw instances through dashboard does not work and is extremely hard to troubleshoot

Added by Christian Huebner about 1 year ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
Tags:
dashboard rgw ssl radosgw quincy
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

I am trying to deploy RadosGW instance with SSL capabilities through the dashboard for test purposes. Deploying non-SSL RadosGWs works fine. When I try RadosGWs, with a certificate in PEM form either pasted in or loaded from a file, all RadosGW instances fail with a red 'error' marker. No error messages are available, neither in the Daemon Events box ("2 minutes ago - daemon:rgw.ssl.ceph2.juxksa Deployed rgw.ssl.ceph2.juxksa on host 'ceph2'"), nor in the Service Events tab ('
3 minutes ago - service:rgw.ssl service was created'). Troubleshooting from the command line didn't give me a clear root cause either, as logs for the service do never show up in the Ceph logs location (/var/log/ceph).

I understand this is not a common use case and for production would have chosen an SSL enabled reverse proxy, but this should be handled gracefully by the dashboard. If it is expected to work and doesn't it should give a useful amount of debug information. I could also deploy manually, but the idea is to test the dashboard to see what our customers may see.

Ceph version is a fresh installation of 17.2.5 (current) for aarch86 on Rocky Linux.

Also available in: Atom PDF