Bug #64291: Incorrect OSD transaction id type, flipping on Windows - Ceph - Ceph

Actions

Copy link

Bug #64291

open

Incorrect OSD transaction id type, flipping on Windows

Added by Lucian Petrut 3 months ago. Updated 3 months ago.

Status:

Pending Backport

Priority:

Normal

Assignee:

Lucian Petrut

Category:

Target version:

% Done:

Source:

Tags:

backport_processed

Backport:

reef quincy

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

55413

Crash signature (v1):

Crash signature (v2):

Description

The MOSDOp constructor receives the the transaction ID as a long instead of ceph_tid_t.

The issue is that "long" uses 32b on Windows instead of 64 bits, so it flips after about 2 billion requests. At that point, the OSD replies are dropped because of transaction ID mismatches.

We'll solve the issue by using the correct type for the transaction id, specifically ceph_tid_t.

```
2024-02-01T13:43:28.420+0000 24 15 client.23819.objecter _send_op 2147492334 to 4.e on osd.2
2024-02-01T13:31:07.044+0000 24 1 -- 13.13.13.10:0/1810840372 --> [v2:13.13.13.13:6818/861966396,v1:13.13.13.13:6819/861966396] -- osd_op(unknown.0.0:18446744071562076507 4.e
...
2024-02-01T13:31:07.044+0000 d 7 client.23819.objecter handle_osd_op_reply 18446744071562076507 ondisk ... stray
```

Related issues 2 (0 open — 2 closed)