[SOLVED] xfs_repair and segmentation fault error

circus78 · 08-24-2017, 02:24 AM

Hi,
I need to check a 14 TB xfs volume.
When I use xfs_check /dev/vda3, I get this error:

Code:

dir 2664293 block 8388612 extra leaf entry 57df663a 10c2
dir ino 2664293 missing leaf entry for 8c750575/10c2

(after a while, I get "Killed" message).

I tried with xfs_repair -n /dev/vda3, and I get "segmentation fault":

Code:

Phase 1 - find and verify superblock...
Phase 2 - using internal log
        - scan filesystem freespace and inode maps...
        - found root inode chunk
Phase 3 - for each AG...
        - scan (but don't clear) agi unlinked lists...
        - process known inodes and perform inode discovery...
        - agno = 0
doubling cache size to 94160
doubling cache size to 188320
entry at block 8 offset 1552 in directory inode 2664293 has illegal name "would clear entry
        - agno = 1
        - agno = 2
        - agno = 3
        - agno = 4
        - agno = 5
        - agno = 6
        - agno = 7
        - agno = 8
        - agno = 9
        - agno = 10
        - agno = 11
        - agno = 12
        - agno = 13
        - process newly discovered inodes...
Bad flags set in inode 12926593676inode 12926593684 has RT flag set but there is no RT deviceinode 12926593684 not rt bitmapdirectory flags set on non-directory inode 12926593684Phase 4 - check for duplicate blocks...
        - setting up duplicate extent list...
        - check for inodes claiming duplicate blocks...
        - agno = 0
entry " would clear inode number in entry at offset 1552...
doubling cache size to 376640
doubling cache size to 753280
Segmentation fault

How can I handle this?

This system is a VM with 3 GB RAM, Debian 7.11
Thank you very much

business_kid · 08-24-2017, 03:39 AM

14TB is a huge disk. I worry that registers held may simply not be big enough in xfs_check.

That aside, a segmentation fault is a memory page error - addressing a memory page that doesn't exist or other such memory error. Do the basics and check memory, reinstall the xfs repair software or run it from another drive. Then run it with --verbose options and pipe it to a log. You can add it to your bug report.

circus78 · 08-24-2017, 08:57 AM

Quote:

Originally Posted by business_kid

14TB is a huge disk. I worry that registers held may simply not be big enough in xfs_check.

That aside, a segmentation fault is a memory page error - addressing a memory page that doesn't exist or other such memory error. Do the basics and check memory, reinstall the xfs repair software or run it from another drive. Then run it with --verbose options and pipe it to a log. You can add it to your bug report.

I just added 2 GB RAM to this machine, and xfs_repair -n completed.

business_kid · 08-24-2017, 09:15 AM

Great. Mark it solved then.

rknichols · 08-24-2017, 05:08 PM

You should also file a bug report. A program should not segfault just because it can't get the amount of memory it wanted.