Project

General

Profile

Actions
Copy link

Bug #74590

open

s3_user_quota-run.sh: Can't connect to trial026..:80 (Name or service not known)

Added by Casey Bodley about 2 months ago. Updated 17 days ago.

Status:
New
Priority:
Urgent
Assignee:
Adam Emerson
Target version:
-
% Done:

0%

Source:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Tags (freeform):
Merge Commit:
Fixed In:
Released In:
Upkeep Timestamp:

Description

from https://qa-proxy.ceph.com/teuthology/ivancich-2026-01-22_18:54:41-rgw-main-distro-default-trial/12997/teuthology.log

2026-01-22T19:16:29.426 INFO:tasks.workunit.client.0.trial026.stderr:+ exec perl /home/ubuntu/cephtest/clone.client.0/qa/workunits/rgw/s3_user_quota.pl
2026-01-22T19:17:32.659 INFO:tasks.workunit.client.0.trial026.stderr:Can't connect to trial026..:80 (Name or service not known)
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:Name or service not known at /usr/share/perl5/LWP/Protocol/http.pm line 50.
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:bucket buck-2026-00-22-19-16-29 create failed
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:Can't connect to trial026..:80 (Name or service not known)
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:
2026-01-22T19:17:32.660 INFO:tasks.workunit.client.0.trial026.stderr:Name or service not known at /usr/share/perl5/LWP/Protocol/http.pm line 50.
2026-01-22T19:17:32.665 DEBUG:teuthology.orchestra.run:got remote process result: 22

Actions
Copy link
 #1

Updated by J. Eric Ivancich about 2 months ago

  • Assignee set to Adam Emerson
Actions
Copy link
 #2

Updated by Casey Bodley 21 days ago

discussed in slack thread https://ceph-storage.slack.com/archives/C1HFJ4VTN/p1772446578424179

Ilya Dryomov: Given that it's on Ubuntu, it could be that there is still something wrong with /etc/hostname such that Perl's hostfqdn() gets confused (edited) 
Ilya Dryomov: I reported something similar in the context of iSCSI and @David Galloway put up https://github.com/ceph/ceph-cm-ansible/pull/821 to address that
Ilya Dryomov: I was under the impression that some version of that PR got deployed despite it not being merged, but I'm not certain
Ilya Dryomov: @Dan Mick would know more
Dan Mick: merged it.  LMK if this still seems to be an issue

Actions
Copy link
 #3

Updated by Ilya Dryomov 17 days ago

@Dan Mick Despite you merging https://github.com/ceph/ceph-cm-ansible/pull/821, the issue still persists
I'm not sure about the RGW test that Nithya started this thread with, but the basic problem is the following
On 10.20.193.21:
$ python3 -c "import socket; print(socket.getfqdn('10.20.193.21'))"
trial021
$ python3 -c "import socket; print(socket.getfqdn('10.20.193.52'))"
trial052.front.sepia.ceph.com
$ cat /etc/hostname
trial021
Because of how /etc/hostname is set up, the FQDN doesn't get returned on the node for the node itself
This breaks tests that rely on hostnames and do cross-checks between two or more nodes
Both Ubuntu 22.04 and CentOS Stream 9 jobs are affected the same way

Actions
Copy link

Also available in: Atom PDF