On Thu, Jul 18, 2013 at 01:04:37PM +0200, Andreas Klauer wrote:
> Hi,
>
> having some problems growing XFS, using Linux 3.10.1 x86_64, xfsprogs 3.1.11.
>
> The filesystem is on LVM, and I made the volume larger. I did this several
> times before and it never causes any issue.
>
> This time, trying to grow the filesystem failed and it seemed corrupt.
> xfs_repair restored its original size somehow (even though shrinking is
> not supposed to be possible?). Unfortunately I did not save the output
> of the commands the first time around, but the issue is reproducable
> in its current state.
>
> If anyone could tell me what I'm doing wrong?
>
> # mount
> > /dev/mapper/lvm-xfs on /mnt/xfs type xfs (rw)
>
> # df -h
> > /dev/mapper/lvm-xfs 2.0T 1.9T 102G 95% /mnt/xfs
>
> # blockdev --getsize64 /dev/lvm/xfs
> > 2306397437952
>
> # xfs_info /mnt/xfs
> > meta-data=/dev/mapper/lvm-xfs isize=256 agcount=16, agsize=32768000
> > blks
> > = sectsz=4096 attr=2
> > data = bsize=4096 blocks=524288000, imaxpct=25
> > = sunit=0 swidth=0 blks
> > naming =version 2 bsize=4096 ascii-ci=0
> > log =internal bsize=4096 blocks=32768, version=2
> > = sectsz=4096 sunit=1 blks, lazy-count=1
> > realtime =none extsz=4096 blocks=0, rtextents=0
>
> # xfs_growfs /mnt/xfs
> > meta-data=/dev/mapper/lvm-xfs isize=256 agcount=16, agsize=32768000
> > blks
> > = sectsz=4096 attr=2
> > data = bsize=4096 blocks=524288000, imaxpct=25
> > = sunit=0 swidth=0 blks
> > naming =version 2 bsize=4096 ascii-ci=0
> > log =internal bsize=4096 blocks=32768, version=2
> > = sectsz=4096 sunit=1 blks, lazy-count=1
> > realtime =none extsz=4096 blocks=0, rtextents=0
> > xfs_growfs: XFS_IOC_FSGROWFSDATA xfsctl failed: Structure needs cleaning
> > data blocks changed from 524288000 to 563085312
What's in dmesg?
> > Phase 2 - using internal log
> > - zero log...
> > - scan filesystem freespace and inode maps...
> > primary/secondary superblock 4 conflict - AG superblock geometry info
> > conflicts with filesystem geometry
> > reset bad sb for ag 4
> > primary/secondary superblock 2 conflict - AG superblock geometry info
> > conflicts with filesystem geometry
> > reset bad sb for ag 2
> > primary/secondary superblock 1 conflict - AG superblock geometry info
> > conflicts with filesystem geometry
> > reset bad sb for ag 1
> > primary/secondary superblock 3 conflict - AG superblock geometry info
> > conflicts with filesystem geometry
> > reset bad sb for ag 3
> > sb_icount 0, counted 8704
> > sb_ifree 0, counted 854
> > sb_fdblocks 0, counted 26540703
So it looks like it got to AG 5 and failed for some reason....
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
|