[Ocfs2-users] Ghost files in OCFS2 filesystem

[Sunil Mushran]

2.6.31-rc1.

Brian Kroth wrote:
> I didn't see this in the bug list.  Which mainline release is this fixed
> in?
>
> Thanks,
> Brian
>
> Sunil Mushran <sunil.mushran at oracle.com> 2009-08-20 17:46:
>   
>> Yes, this is a known issue in OCFS2 1.4.1 and 1.4.2. That is assuming
>> no process in the cluster has that file open. We have the fix. It will be
>> available with 1.4.3 which is in testing.
>>
>> This was discussed in the email announcing the 1.4.2 release.
>> ====================================================================
>> http://oss.oracle.com/pipermail/ocfs2-announce/2009-June/000028.html
>>
>> 1. Oracle# 7643059 - Orphan files not getting deleted
>> When one unlinks a file, its inode is moved to the orphan directory and
>> deleted when it is no longer in-use across the cluster. As part of the
>> scheme, the node that unlinks the file, informs interested nodes of the
>> same, asking the last node to stop using that inode to recover the space
>> allocated to it. However, this scheme fails if memory pressure forces a
>> node to forget to delete the inode on close.
>> This issue was introduced in OCFS2 1.4.1. While we have fixed this issue,
>> the fix did not make it into this release. Users running into this issue
>> can call Oracle Support and ask for an interim release with this fix.
>> ====================================================================
>>
>> Workaround:
>> If on 1.4.2, mount the fs on another node. Chances are it will delete
>> the orphans. If not or if you are on 1.4.1, umount vol on all nodes and
>> run fsck.ocfs2 -f.
>>
>> You could ping support to get an interim fix. But we are close to releasing
>> 1.4.3. So maybe better if you wait for that.
>>
>> Sunil
>>
>>
>> Shave, Chris wrote:
>>     
>>> Hi,
>>>  
>>> I have encountered an issue on an Oracle RAC cluster using ocfs2, OS 
>>> is RH Linux 5.3.
>>>  
>>> One of the ocfs2 filesystems appears to be 97% full, yet when I look 
>>> at the files in there they only equal about 13gig (filesystems is 
>>> 40gig in size).
>>>  
>>> I have seen this sort of thing in HP-UX but that involved a process 
>>> who's output file was deleted but the process hadn't been stopped 
>>> properly, once we killed the offending process the space was released, 
>>> but I can't seem to find any process on this Linux server that is 
>>> using or writing files to that filesystem.
>>>  
>>> File system:
>>>  
>>> Filesystem            Size  Used Avail Use% Mounted on
>>> /dev/emcpoweri1        40G   39G  1.9G  96% /oraexport
>>> File listing: Other than directories, these are the only files in that 
>>> filesystem, nothing in lost+found either..
>>>  
>>> [root at aumel21db01cn01]# ll
>>> total 11926784
>>> -rw-rw---- 1 oracle oinstall 12210978816 Aug 21 05:23   
>>> 00000.Full.090821.dmp
>>> -rw-rw-r-- 1 oracle oinstall     1920327 Aug 21 05:23      
>>> 00000.Full.090821.log
>>> Any assistance with what is going on would be greatly appreciated.
>>>       
>> _______________________________________________
>> Ocfs2-users mailing list
>> Ocfs2-users at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>>