xfs_repair has been running for a very long time, no disk IO.

xfs_repair has been running for a very long time, no disk IO.

[Emmanuel Florac]

On a quite badly mangled filesystem, xfs_repair (v. 3.1.4) seems to
never end. It has run for 16 hours with absolutely no disk activity.

Usually xfs_repair takes much less than one hour on a similar (9 TB)
filesystem.

I know I should try the latest version. What else can be wrong? 
What should I try?

Last lines of output of the still running xfs_repair where it has 
been stuck since yesterday night:

entry "Panel-2333" at block 3 offset 2464 in directory inode 2190114192 references non-existent inode 60238299
        clearing inode number in entry at offset 2464...
entry "Panel-2316" at block 3 offset 2512 in directory inode 2190114192 references non-existent inode 60238300
        clearing inode number in entry at offset 2512...
entry "Panel-2309" at block 3 offset 2560 in directory inode 2190114192 references non-existent inode 60238301
        clearing inode number in entry at offset 2560...
entry "Panel-2292" at block 3 offset 2608 in directory inode 2190114192 references non-existent inode 60238302
        clearing inode number in entry at offset 2608...
entry "Panel-2290" at block 3 offset 2656 in directory inode 2190114192 references non-existent inode 60238303
        clearing inode number in entry at offset 2656...
bad attribute format 1 in inode 2190120814, resetting value
        - agno = 2
        - agno = 3
        - agno = 4
        - agno = 5
        - agno = 6
        - agno = 7
        - agno = 8
        - agno = 9
        - agno = 10
        - agno = 11
        - agno = 12
        - agno = 13
        - agno = 14
        - agno = 15
        - agno = 16
        - agno = 17
        - agno = 18
        - agno = 19
        - agno = 20
        - agno = 21
        - agno = 22
        - agno = 23
        - agno = 24
        - agno = 25
        - agno = 26
        - agno = 27
        - agno = 28
        - agno = 29
        - agno = 30
        - agno = 31
Phase 5 - rebuild AG headers and trees...
        - reset superblock...
Phase 6 - check inode connectivity...
        - resetting contents of realtime bitmap and summary inodes
        - traversing filesystem ...



-- 
------------------------------------------------------------------------
Emmanuel Florac     |   Direction technique
                    |   Intellique
                    |	<eflorac@intellique.com>
                    |   +33 1 78 94 84 02
------------------------------------------------------------------------

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: xfs_repair has been running for a very long time, no disk IO.

[Dave Chinner]

On Tue, Apr 09, 2013 at 12:17:26PM +0200, Emmanuel Florac wrote:
> 
> On a quite badly mangled filesystem, xfs_repair (v. 3.1.4) seems to
> never end. It has run for 16 hours with absolutely no disk activity.
> 
> Usually xfs_repair takes much less than one hour on a similar (9 TB)
> filesystem.
> 
> I know I should try the latest version. What else can be wrong? 
> What should I try?

Either use the latest version, or use the -P option.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: xfs_repair has been running for a very long time, no disk IO.

[Emmanuel Florac]

Le Tue, 9 Apr 2013 22:02:36 +1000
Dave Chinner <david@fromorbit.com> écrivait:

> Either use the latest version, or use the -P option.
> 

Currently running the latest version (3.1.11-pre2). I'll wait for a
while (it takes something like 10 minutes to reach the point at which it
freezes over). I suppose -P makes it slower?

-- 
------------------------------------------------------------------------
Emmanuel Florac     |   Direction technique
                    |   Intellique
                    |	<eflorac@intellique.com>
                    |   +33 1 78 94 84 02
------------------------------------------------------------------------

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: xfs_repair has been running for a very long time, no disk IO.

[Dave Chinner]

On Tue, Apr 09, 2013 at 02:30:10PM +0200, Emmanuel Florac wrote:
> Le Tue, 9 Apr 2013 22:02:36 +1000
> Dave Chinner <david@fromorbit.com> écrivait:
> 
> > Either use the latest version, or use the -P option.
> > 
> 
> Currently running the latest version (3.1.11-pre2). I'll wait for a
> while (it takes something like 10 minutes to reach the point at which it
> freezes over). I suppose -P makes it slower?

Yes, it turns prefetching off, which is what usually what causes the
hangs. Though 3.1.4 is pretty old, and I think a few prefetch hangs
were fixed in 3.1.8....

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: xfs_repair has been running for a very long time, no disk IO.

[Emmanuel Florac]

Le Tue, 9 Apr 2013 22:54:02 +1000
Dave Chinner <david@fromorbit.com> écrivait:

> Yes, it turns prefetching off, which is what usually what causes the
> hangs. Though 3.1.4 is pretty old, and I think a few prefetch hangs
> were fixed in 3.1.8....
> 

Actually 3.1.11-pre2 did it without -P. Thanks.

-- 
------------------------------------------------------------------------
Emmanuel Florac     |   Direction technique
                    |   Intellique
                    |	<eflorac@intellique.com>
                    |   +33 1 78 94 84 02
------------------------------------------------------------------------

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs