Project

General

Profile

Actions

Bug #13676

open

teuthology-nuke does not clear all entries in /var/run/ceph

Added by Anonymous over 8 years ago. Updated over 8 years ago.

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

0%

Source:
Q/A
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Crash signature (v1):
Crash signature (v2):

Description

I ran teuthology using the following yaml. Note that this does not have a ceph: clause. Instead it uses calamari-setup to install a ceph cluster.

roles:
- [mon.a, mon.c, osd.0, osd.1, osd.2, client.0]
- [mon.b, osd.3, osd.4, osd.5, client.1]
machine_type: vps
overrides:
  ceph:
    log-whitelist:
    - wrongly marked me down
tasks:
- ssh_keys:
- calamari_setup:
    test_image: http://file.rdu.redhat.com/~kdreyer/download.inktank.com/enterprise/1.2.3/rhceph-1.2
.3-ubuntu-14.04-x86_64.iso
    email: wusui@redhat.com
- install.ship_utilities: null
- interactive:
- workunit:
    clients:
      all:
        - rados/load-gen-mostlyread.sh
targets:
  vpm049.front.sepia.ceph.com: ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCz9GfRQpdIgPXEwA6R4QxC/B4h0mOc3
F+uizZKSiesUA451n2KpoUHr+0YgkrJSJTh2LpbyQg93MK8UCWgGfo3BPAOEfbCIa2im5/CwayaYafqOlJ7wvifinKjDnGJzD7Mj
90T8rVk+KK/+zyYU7qwhQpEQAOzVwdsPDKc3iNZSILAvJE7QlO851b2WMZRoDKC3vtgJOssWnLHLEcjfiN79AkRTjN2eSUijTME6
PvVElnWKOBGZbY69cIXjZ0iTh0IR6ZsOFicjOB4P7tDnbjhy3N2I42ZWfHgSnjmrlLLB6qkUNpOUQK7p3jhkh04KcoEQbZMXNC/b
UvixHposRAl
  vpm140.front.sepia.ceph.com: ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDsnEduKlpjUC9M/uTlPL2/J5PCzuiVQ
9hmwd2B2j17Kowd5cTeFZ73GgzKSjTRekFEPK+/7D8RnbmftOvKo9iUTdAftSXeYNjzcdAAZ/dlDkHwUOwtLlOk8huHFwElpIsrQ
ryr9ZyyK4A2HwDravb5R7+W+gazW0wi2v54WxQj4/hqBZRurp8SWJNsnr2qcCK84iejw5EHJvrAYYjwzOsFHWFqxiKbPv7rkeLxm
FTwMT0PD0bRYBaaUnF7sswmMCGZaXw8le1dmCAPxT4dlymgWLmc5PA3YhbZD7Qu06FVEGCkOiTrWmCl3Pp0hJsvQAjAyYk3BF1mP
CtCGL/vsTYZ

At the interactive point, ceph appears to be up (run ceph from another window).

Once I proceed from here, the workunit fails due to a keyring issue (I am not reporting this yet because it still may be a problem in my setup).

I then run teuthology-nuke. It seems to complete okay, but afterwards /var/run/ceph contain socket entries that caused "File already exists" errors the next time that I ran teuthology.

ubuntu@vpm049:~$ ls -ls /var/run/ceph
total 0
0 srwxr-xr-x 1 root root 0 Nov  3 00:52 ceph-osd.1.asok

ubuntu@vpm140:~$ ls -ls /var/run/ceph
total 0
0 srwxr-xr-x 1 root root 0 Nov  3 00:52 ceph-osd.2.asok
0 srwxr-xr-x 1 root root 0 Nov  3 00:52 ceph-osd.4.asok
0 srwxr-xr-x 1 root root 0 Nov  3 00:53 ceph-osd.5.asok

I suppose the use of calamari_setup and lack of ceph would be a configuration that has not been tried yet. Still, I would expect all of these sockets to be deleted by the nuke command.

Actions #1

Updated by Anonymous over 8 years ago

  • Subject changed from teruthology-nuke does not clear all entries in /var/run/ceph to teuthology-nuke does not clear all entries in /var/run/ceph
Actions #2

Updated by Anonymous over 8 years ago

  • Source changed from other to Q/A
Actions

Also available in: Atom PDF