[Ocfs2-users] Reproducing fragmentation and out of space error

[Srinivas Eeda]

Nick Anderson wrote:
> On 05/31/2012 08:25 AM, Nick Anderson wrote:
>   
>> Is there some specific output from debugfs.ocfs that will tell me when I
>> am approaching the situation?
>>     
>
>  From an old bug report
>   
>> http://oss.oracle.com/bugzilla/show_bug.cgi?id=915#c10
>> http://oss.oracle.com/bugzilla/show_bug.cgi?id=915#c24
>>     
>
> I used the defaults when I created this filesystem.
> debugfs.ocfs2 -R "stat //inode_alloc:0000" /dev/mapper/cm_shared_ocfs2 | 
> grep "Clusters per"
>          Clusters per Group: 1024   Bits per Cluster: 1
>
> I thought that I could look at the global_bitmap and look at the Contig 
> column and any time there were no values in that column that were 
> greater than or equal to 1024 I should be seeing the error. But that 
> cant be the case, right now I don't have any contig values over 400 and 
> I am not getting the out of space error.
>
> I must be mis-understanding something. Can someone clarify what I can 
> look at to see the situation approach and know when I am in the state 
> that produces the out of space error behavior?
>   
What kernel version and ocfs2 version are you using?

If you are running into enospc even though you see free space, then it's 
most likely because of fragmentation. metadata files inode_alloc and 
extent_alloc are full and fragmentation prevents them from growing