[dm-devel] [dm-thin] metadata corruption

[黃崇智]

Dear All,

I'm currently doing pressure testing for dm-thin on 'Linux 3.13.5, Ubuntu
13.10' and finally encounter the issue with metadata corruption.

Using following steps:
1. Creating a pool with metadata device 15GB, data device 3.6TB
2. Creating 60 thin volumes(thin0 ~ thin59) each with 60GB
3. Issuing I/O to thin0~thin9 continuously (e.q. dd if=/dev/zero
of=/mnt/thin0 bs=1M  count=1024)
4. Taking snapshot for all the thin volumes every hour

After about 3 to 20 hours,
I found out all the issued I/Os are stopped, and kernel message shows up
following content:
*"device-mapper: space map common: unable to decrement a reference count
below 0"*
*"device-mapper: thin: 252:0: dm_thin_insert_block() failed: error = -22"*
*" device-mapper: thin: 252:0: switching pool to read-only mode"*

Is any way to fix this problem?

PS:
Attached files are the kernel message, and the python script to reproduce
such issue


Any help would be grateful.
Thanks,

Webber
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20140226/5e1ef201/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: generic_run.py
Type: application/octet-stream
Size: 3393 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20140226/5e1ef201/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kern.log
Type: application/octet-stream
Size: 1520213 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20140226/5e1ef201/attachment-0001.obj>