[OpenAFS-devel] CopyOnWrite failure: ihandle.c Patch: Code Cleanup

[hoffman@cs.pitt.edu]

> unfortunately the ihandle patch does not solve the CopyOnWrite
> problem. We had another CopyOnWrite failure this morning ...

We had one last night, 51 seconds after the backup volume was recloned:

VolserLog:
Thu Apr  4 15:49:58 2002 1 Volser: Clone: Recloning volume 536874136 to volume 536874138

FileLog:
Thu Apr  4 15:50:49 2002 CopyOnWrite failed: volume 536874136 in partition /vicepb  (tried reading 4096, read 0, wrote 0, errno 11) volume needs salvage

> Because only the .kde and .netscape directories we corrupted 
> this time I did not restore the whole volume from backup but 
> did a "salvager ... -orphan attach" twice. When I moved the 
> volume away from this file server afterwards I got lots of 
> 
> Fri Apr  5 13:38:20 2002 ReallyRead(): read failed device 0 inode
> 80BD460 errno 5
> 
> in the file server log. 

We didn't get any of those.  I was able to get a volinfo snapshot of
the corrupted volume before and after salvaging.  The files are:

ftp://ftp.cs.pitt.edu/hoffman/openafs/SalvageLog.536874136
ftp://ftp.cs.pitt.edu/hoffman/openafs/volinfo.536874136
ftp://ftp.cs.pitt.edu/hoffman/openafs/volinfo.536874136.after_salvage

The corrupted vnode (2945) was not the top-level vnode this time.
It was, however, under heavy use.

	---Bob.
--
Bob Hoffman, N3CVL      University of Pittsburgh           Tel: +1 412 624 8404
hoffman@cs.pitt.edu     Department of Computer Science     Fax: +1 412 624 8854