On October 21 at 22:14 Patrik Jonsson wrote:
> Hi all,
>
> I have an xfs file system which has been running fine for quite some
> time, but on which I now have a media error. I'm trying to track down
> which file is hit. What I've done so far is:
>
> xfs_db> blockget -n -i 393842234
> upinode 393842234 add link, now 1
> inode 393842234 mode 0100644 fmt extents afmt extents nex 1 anex 0 nblk
> 445 sz 1821919
> inode 393842234 nlink 1 not dir
> inode 393842234 extent [0,56060208,445,0]
> setting inode to 393842234 for block 3/5728560
> ...
> and many more such.
>
> The plan was to then do ncheck -i 393842234 to get the file name, but at
> this point xfs_db has been running for 4 hours, is using 5.6GB of memory
> and since the machine only has 2GB, it's swapping like crazy. Is this
> normal or is something wrong?
Not sure, but this sounds kinda sketchy. Someone else will need to comment
on the behavior of xfs_db.
> I just want to know the name of a file...
Try 'find /data -inum 393842234' to get you moving forward again until we
understand what's going on with xfs_db.
> I'm running a 2.6.11.12 kernel, and xfs_info provides the following, in
> case it's any help:
>
> [root@localhost ~]# xfs_info /data
> meta-data=/data isize=256 agcount=32,
> agsize=13326352 blks
> = sectsz=512
> data = bsize=4096 blocks=426443232, imaxpct=25
> = sunit=16 swidth=112 blks, unwritten=1
> naming =version 2 bsize=4096
> log =internal bsize=4096 blocks=32768, version=1
> = sectsz=512 sunit=0 blks
> realtime =none extsz=458752 blocks=0, rtextents=0
>
>
> Any help would be appreciated. Thanks,
>
> /Patrik Jonsson
|