Bug #2105: filestore: mkfs does not create initial snap - Ceph - Ceph

Actions

Copy link

Bug #2105

closed

filestore: mkfs does not create initial snap

Added by Yunpeng Gao about 12 years ago. Updated about 12 years ago.

Status:

Resolved

Priority:

Urgent

Assignee:

Sage Weil

Category:

OSD

Target version:

v0.44

% Done:

Source:

Community (user)

Tags:

Backport:

Regression:

Severity:

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

This bug almost the same as this bug:http://tracker.newdream.net/issues/1707
I followed the instruction:http://ceph.newdream.net/docs/latest/ops/install/mkcephfs/ to install ceph and configure the cluster on Debian Squeeze 6.0.3 on all nodes.
After I run the command:"mkcephfs - a - c mycluster.conf - k mycluster.keyring"
I did not get any errors, but when I run the command:"ceph - k mycluster.keyring -c mycluster.conf health".
The result is:

2012-02-27 03:20:58.064745 mon < - [health]
2012-02-27 03:20:58.065714 mon.0 -> 'HEALTH_ERR no osds' (0)

Then I check the log, found this error:

_2012-02-27 00:47:05.909476 7f0679f73780 filestore(/ceph/osd.0) error (17) File exists not handled on operation 20 (op num 1, counting from 1)
2012-02-27 00:47:05.909485 7f0679f73780 filestore(/ceph/osd.0) unexpected error code
2012-02-27 00:47:05.909503 7f0679f73780 filestore(/ceph/osd.0) transaction dump:
1: mkcoll meta
2: write meta 23c2fcde/osd_superblock/0 0~106 (106)
3: mkcoll temp

os/FileStore.cc: In function 'unsigned int FileStore::_do_transaction(ObjectStore::Transaction&, uint64_t)' thread 7f0679f73780 time 2012-02-27 00:47:05.909526
os/FileStore.cc: 2475: FAILED assert(0 == "unexpected error")
ceph version 0.42.2 (commit:732f3ec94e39d458230b7728b2a936d431e19322)_

the full log is here:

Files

osd.0.log (6.61 KB) osd.0.log

Yunpeng Gao, 02/26/2012 11:19 AM

Actions

Copy link

Updated by Sage Weil about 12 years ago

Category set to OSD
Priority changed from Normal to Urgent
Source changed from Development to Community (user)

Actions

Copy link

Updated by Sage Weil about 12 years ago

Subject changed from After ceph cluster installed, OSD initialization fails with: error error 17: error (17) File exists not handled on operation 20 (op num 1, counting from 1) to filestore: mkfs does not create initial snap
Status changed from New to 7
Assignee set to Sage Weil
Target version set to v0.44

Thanks for the report!

The mkfs doesn't create an initial snap, so if we crash/stop before creating one, our first journal events will get replayed against a dirty current/. See wip-2105 for fix.

Actions

Copy link