http://bugzilla.kernel.org/show_bug.cgi?id=10421
------- Comment #39 from sandeen@xxxxxxxxxx 2008-05-08 06:58 -------
I might be hitting something similar here. It's an odd setup, 2.6.25 xfs
backported to an older kernel; I need to try upstream. But if I run this test
overnight...
#!/bin/bash
mkdir /mnt/test/dir1
mkdir /mnt/test/dir2
mkdir /mnt/test/dir3
mkdir /mnt/test/dir4
dd if=/dev/zero of=/mnt/test/dir1/file1 bs=1M count=20480 &
dd if=/dev/zero of=/mnt/test/dir2/file2 bs=1M count=20480 &
dd if=/dev/zero of=/mnt/test/dir3/file3 bs=1M count=20480 &
dd if=/dev/zero of=/mnt/test/dir4/file4 bs=1M count=20480 &
sync
sync
rm -f /mnt/test/dir1/file1 &
rm -f /mnt/test/dir2/file2 &
rm -f /mnt/test/dir3/file3 &
rm -f /mnt/test/dir4/file4 &
wait
I eventually hit something similar. I changed NULLAGINO and NULLFSINO:
-#define NULLFSINO ((xfs_ino_t)-1)
-#define NULLAGINO ((xfs_agino_t)-1)
+#define NULLFSINO ((xfs_ino_t)-2)
+#define NULLAGINO ((xfs_agino_t)-3)
in both userspace & kernelspace and when I trip I get:
xfs_inotobp: xfs_imap() returned an error 22 on sdb5 for inode
4294967293. Returning error.
so that is a NULLAGINO
-Eric
--
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
|