Hi Steve
Steve Lord [lord@xxxxxxx] wrote:
>
> I think you got bitten by stripe alignment. Inode clusters are allocated
> on stripe boundaries. You probably have no boundaries left free, so
> it cannot allocate any inode space.
>
> If you go into xfs_db again, and run freesp, then run frag, what does
> it say? We probably need to revisit how files are getting allocated
> for NFS, I think it is not doing a very good job in this case. What
> sort of file size are you talking about here, the numbers say about 175K
> but I want to check.
Should i run xfs_db with freesp and frag?
I had run frag before the no space left situation (about 70% usage i guess),
it reports 14% fragmentation.
The files are actually copied from our main fileserver. User Homes,
installed appications und a lot of CAE data (100+MB) and a few multigigabyte
files. So you have everything from 0 bytes to a few GB.
>
> In this setup fsr will not do anything, and it can sometimes have the
> effect of defragmenting files, but fragmenting the remaining free space.
I run the fsr when the filesystem wasn't filled up so much. I had done this
mostly for stressing. If i remember correctly there was an issue running fsr
on nfs exports.
btw: i have tried to trash the fs hard. power off while lots of different
access patterns. Only one time xfs 1.2 hangs after log replay, 1.3pre2
works everytime. And i got not a single fs corruption.
utz
|