Skip to contentSkip to conversation

xfsp_repair segfault (3.1.4 & 3.1.6)

DISCUSSION nanashi 2 participants 2 messages Oct 28, 2011 – Nov 4, 2011

lore
Share & tools
lore

nanashi

author
#0 lore raw

Hi,

I have a corrupted RAID5 xfs filesystem from an intel SS-4000 NAS storage. I'm using slackware-current with 2.6.39.3-smp kernel (32bit & 64bit) When i try xfs_repair I get

xfs_repair: dir2.c:2133: process_dir2: Assertion `(ino !=mp->m_sb.sb_rootino && ino != *parent) || (ino == mp->m_sb.sb_rootino && (ino == *parent || need_root_dotdot == 1))'

I try both x86 and x86_64, I build 3.1.4 and 3.1.6 with DEBUG=-DNDEBUG and both continue but gave me a segfault

[23978.718305] xfs_repair[25800]: segfault at 7fffa1d81ff0 ip 00007f15c1852049 sp 00007fffa257f048 error 6 in libc-2.13.so[7f15c1715000+19b000]

I try xfs_metadump and I get segfault too.

The partition is 2.2TB and I don't have enough space to dd it to an image.

I attach the xfs_repair output before the segfault.

any help is appreciated.

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#1 lore raw
Quoted 14 lines · nanashi

Hi,

I have a corrupted RAID5 xfs filesystem from an intel SS-4000 NAS storage. I'm using slackware-current with 2.6.39.3-smp kernel (32bit & 64bit) When i try xfs_repair I get

xfs_repair: dir2.c:2133: process_dir2: Assertion `(ino !=mp->m_sb.sb_rootino && ino != *parent) || (ino == mp->m_sb.sb_rootino && (ino == *parent || need_root_dotdot == 1))'

I try both x86 and x86_64, I build 3.1.4 and 3.1.6 with DEBUG=-DNDEBUG and both continue but gave me a segfault

[23978.718305] xfs_repair[25800]: segfault at 7fffa1d81ff0 ip 00007f15c1852049 sp 00007fffa257f048 error 6 in libc-2.13.so[7f15c1715000+19b000]
Quoted 8 lines

I try xfs_metadump and I get segfault too.

The partition is 2.2TB and I don't have enough space to dd it to an image.

I attach the xfs_repair output before the segfault.

any help is appreciated.

A corefile + debug binary would be helpful too...

Argh, if metadump segfaults, that will make things tough.

At least with a core + binary we could see just where it blew up.

-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs