[OpenAFS] 1.4.2fc3 and vos move

[John Koyle]

Christopher D. Clausen wrote:
> John Koyle <jkoyle@koyle.org> wrote:
>   
>> I've been running a small single server cell on 1.4.0-rc4 for some
>> time and am just now adding a second server to the mix.  I installed
>> 1.4.2fc2 on the second machine and successfully moved several volumes
>> to it with no problems.
>>     
>
> What OS are you running these on?
>   
This is on Debian stable with 2.6.16-18 kernel from backports.org.  AFS 
is self compiled on the box using: ./configure 
--with-afs-sysname=i386_linux26 --enable-transarc-paths and gcc 3.3.5.

>> Last week when fc3 came out I upgraded the second machine from fc2 to
>> fc3.  Migrating additional volumes now seems to work, but the process
>> fails at the end when it tries to clean up the source:
>>
>> user@host:~$ vos move home.user server1 /vicepa server2 /vicepa
>>
>> Failed to create a transaction on the source volume 536870984
>>   VOLSER: volume is busy
>> vos move: operation interrupted, cleanup in progress...
>> clear transaction contexts
>> move incomplete - attempt cleanup of target partition - no guarantee
>> cleanup complete - user verify desired result
>>
>> This leaves server1 in a bit of disarray as well with **** Could not
>> attach volume 536870984 ****
>> in the output of vos listvol.  I believe the contents of this volume
>> are now empty, and have simply removed the volume entry in /vicepa to
>> clean it up (this may be a bad thing though).
>>
>> However, as mentioned above, the volume is fine on server2 and behaves
>> as expected.
>>
>> Did something change between fc2 and fc3 that may have caused this?
>>     
>
> Hmm... I actually had this problem with fc2 as well (on sun4x_510.)  I 
> thought it was fixed in fc3.  I guess I need to do some more testing...
>
> Do you by chance have any errors in the server logs?
>   
Yes, but nothing very verbose at this point.

server 2's FileLog:
Tue Sep 12 22:07:56 2006 fssync: volume 536871110 moved to 561e140a; 
breaking all call backs
Tue Sep 12 22:07:56 2006 fssync: volume 536871110 restored; breaking all 
call backs
Tue Sep 12 22:07:56 2006 VAttachVolume: Failed to open 
/vicepa//V0536871110.vol (errno 2)
Tue Sep 12 22:07:56 2006 VAttachVolume: Error reading namei vol header 
/vicepa//V0536871113.vol; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Error attaching volume 
/vicepa//V0536871113.vol; volume needs salvage; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Error reading namei vol header 
/vicepa//V0536871113.vol; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Error attaching volume 
/vicepa//V0536871113.vol; volume needs salvage; error=101

server 2's VolserLog:
Tue Sep 12 22:07:38 2006 1 Volser: CreateVolume: volume 536871110 
(home.bob) created
Tue Sep 12 22:07:56 2006 1 Volser: Clone: Cloning volume 536871110 to 
new volume 536871113
Tue Sep 12 22:07:56 2006 1 Volser: Delete: volume 536871110 deleted
Tue Sep 12 22:07:56 2006 VAttachVolume: Failed to open 
/vicepa/V0536871112.vol (errno 2)
Tue Sep 12 22:07:56 2006 VAttachVolume: Error reading namei vol header 
/vicepa/V0536871113.vol; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Error attaching volume 
/vicepa/V0536871113.vol; volume needs salvage; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Failed to open 
/vicepa/V0536871112.vol (errno 2)
Tue Sep 12 22:07:56 2006 VAttachVolume: Failed to open 
/vicepa/V0536871110.vol (errno 2)
Tue Sep 12 22:07:56 2006 VAttachVolume: Error reading namei vol header 
/vicepa/V0536871113.vol; error=101
Tue Sep 12 22:07:56 2006 VAttachVolume: Error attaching volume 
/vicepa/V0536871113.vol; volume needs salvage; error=101

Server 1's FileLog (the destination for the move)
Tue Sep 12 22:07:56 2006 fssync: volume 536871110 restored; breaking all 
call backs
Tue Sep 12 22:07:56 2006 fssync: volume 536871110 restored; breaking all 
call backs

Server 1's VolserLog:
Tue Sep 12 22:07:56 2006 VAttachVolume: Failed to open 
/vicepa/V0536871110.vol (errno 2)
Tue Sep 12 22:07:56 2006 1 Volser: CreateVolume: volume 536871110 
(home.bob) created

> Did any instances crash and dump core?
>
>   
No not yet.
> See bug: http://rt.central.org/rt/Ticket/Display.html?id=38566

This sounds similar to the problem I'm having.  If it's intermittent 
that may explain why the moves were working previously?

John