Project

General

Profile

Actions
Copy link

Bug #62862

open

ceph-volume sometimes miscounts the number of logical volumes on shared DB devices

Added by Benoît Knecht 8 months ago.

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

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
53500
Crash signature (v1):
Crash signature (v2):

Description

On some systems, `pvs --readonly -o pv_name,vg_name,lv_name` returns duplicate entries, which leads `ceph-volume` to think there are move LVs on the DB device than there really is, which can prevent `cephadm` from redeploying and OSD after a disk replacement, even though there's enough space on the DB device.

This can also lead to `ceph-volume lvm list` showing duplicate devices in the `[db]` section of some OSDs, such as

```
[db] /dev/ceph-f3e5517c-a784-48d7-a8b2-6e48d711eea8/osd-db-fc8eeb21-9ce0-400f-9c22-5d311992791d

block device              /dev/ceph-f3e5517c-a784-48d7-a8b2-6e48d711eea8/osd-db-fc8eeb21-9ce0-400f-9c22-5d311992791d
      [...]
      devices                   /dev/nvme2n1,/dev/nvme2n1
```

No data to display

Actions
Copy link

Also available in: Atom PDF