Talk:Using XFS instead of ext3 (network performance boost)

From NAS-Central Buffalo - The Linkstation Wiki
Revision as of 03:00, 18 December 2006 by Davy gravy (Talk | contribs) (→‎Conflict between XFS, JFS and an AudioCD app?)

Jump to: navigation, search

==Conflict between XFS, JFS and an AudioCD app?== Davy gravy 22:00, 17 December 2006 (EST)

Posted by davy_gravy - conversation between him & mindbender

a followup after a week or so of testing and use...

some bad news on the jfs (it could be a random problem (not replicable/reproduceable), it could be because of using fsconvert, or it could be a problem w/ jfs - my guess is that it was caused by the conversion - there is perhaps something that we forgot to do/check as a last step - but I have no idea what that could be...)

After a very good week of use, my hda3 jfs partition no longer mounts. Rebooting doesn't help. running jfs_fsck doesn't help. Attempting to mount manually yields a complaint of "mount: Structure needs cleaning", yet jfs_fsck shows this 

jfs_fsck /dev/hda3 
jfs_fsck version 1.1.7, 22-Jul-2004 
processing started: 12/10/2006 10.39.45 
Using default parameter: -p 
The current device is:  /dev/hda3 
Block size in bytes:  4096 
Filesystem size in blocks:  59605166 
**Phase 0 - Replay Journal Log 
Filesystem is clean.


dmesg seems to show that some aspect of the system seems to believe that it should be using XFS: 

EXT3 FS on hda1, internal journal 
JFS: nTxBlock = 988, nTxLock = 7907 
sr0: scsi3-mmc drive: 8x/48x writer cd/rw xa/form2 cdda tray 
Uniform CD-ROM driver Revision: 3.20 
sr 0:0:0:0: Attached scsi CD-ROM sr0 
sr 0:0:0:0: Attached scsi generic sg0 type 5 
sd 1:0:0:0: Attached scsi generic sg1 type 0 
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended 
end_request: I/O error, dev sr0, sector 1397120 
Buffer I/O error on device sr0, logical block 174640 
end_request: I/O error, dev sr0, sector 1397120 
Buffer I/O error on device sr0, logical block 174640 
end_request: I/O error, dev sr0, sector 1397032 
Buffer I/O error on device sr0, logical block 174629 
end_request: I/O error, dev sr0, sector 1397032 
Buffer I/O error on device sr0, logical block 174629 
end_request: I/O error, dev sr0, sector 1397032 
Buffer I/O error on device sr0, logical block 174629 
end_request: I/O error, dev sr0, sector 1396888 
Buffer I/O error on device sr0, logical block 174611 
r8169: eth0: link up 
eth0: 1000Mbps Full-duplex operation. 
NET: Registered protocol family 10 
lo: Disabled Privacy Extensions 
IPv6 over IPv4 tunneling driver 
Capability LSM initialized 
eth0: no IPv6 routers present 
sg_write: data in/out 56/56 bytes for SCSI command 0x12--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 26/26 bytes for SCSI command 0x5a--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
sg_write: data in/out 12/12 bytes for SCSI command 0x43--guessing data in; 
   program cdparanoia not setting count and/or reply_len properly 
SGI XFS with ACLs, security attributes, no debug enabled 
SGI XFS Quota Management subsystem 
XFS mounting filesystem hda3 
Starting XFS recovery on filesystem: hda3 (logdev: internal) 
Filesystem "hda3": xfs_inode_recover: Bad inode magic number, dino ptr = 0xc15a1400, dino bp = 0xc330 
6840, ino = 132 
Filesystem "hda3": XFS internal error xlog_recover_do_inode_trans(1) at line 2364 of file fs/xfs/xfs_log_ 
recover.c.  Caller 0xc9300e10 
Call Trace: 
[C29EF930] [C000C248] show_stack+0x4c/0x194 (unreliable) 
[C29EF970] [C92E9A50] xfs_error_report+0x60/0x64 [xfs] 
[C29EF980] [C9300874] xlog_recover_do_inode_trans+0x72c/0x80c [xfs] 
[C29EF9F0] [C9300E10] xlog_recover_do_trans+0x13c/0x154 [xfs] 
[C29EFA20] [C9300F2C] xlog_recover_commit_trans+0x60/0x7c [xfs] 
[C29EFA40] [C930106C] xlog_recover_process_data+0xf4/0x1c8 [xfs] 
[C29EFA80] [C9301D98] xlog_do_recovery_pass+0x1e8/0x810 [xfs] 
[C29EFB30] [C930243C] xlog_do_log_recovery+0x7c/0xac [xfs] 
[C29EFB60] [C9302484] xlog_do_recover+0x18/0x148 [xfs] 
[C29EFB80] [C9302650] xlog_recover+0x9c/0xd8 [xfs] 
[C29EFBB0] [C92FA6D0] xfs_log_mount+0xa8/0x124 [xfs] 
[C29EFBE0] [C9303B1C] xfs_mountfs+0x5a4/0xaf8 [xfs] 
[C29EFC90] [C92F681C] xfs_ioinit+0x38/0x4c [xfs] 
[C29EFCB0] [C930ABA4] xfs_mount+0x258/0x3e8 [xfs] 
[C29EFCE0] [C9320040] vfs_mount+0x44/0x5c [xfs] 
[C29EFCF0] [C931FE6C] xfs_fs_fill_super+0x9c/0x200 [xfs] 
[C29EFD90] [C00664FC] get_sb_bdev+0x128/0x16c 
[C29EFDE0] [C931FFEC] xfs_fs_get_sb+0x1c/0x2c [xfs] 
[C29EFDF0] [C00667E8] vfs_kern_mount+0xb0/0x158 
[C29EFE20] [C00668D0] do_kern_mount+0x40/0x64 
[C29EFE50] [C007FE4C] do_new_mount+0x94/0xc4 
[C29EFE80] [C0080674] do_mount+0x1ac/0x1f4 
[C29EFF00] [C0080A50] sys_mount+0xa4/0xf4 
[C29EFF40] [C0004100] ret_from_syscall+0x0/0x38 
XFS: log mount/recovery failed: error 117 
XFS: log mount failed 
XFS mounting filesystem hda3


My gut feel is that in the conversion process, something was left out...?

Double-checking /etc/fstab shows jfs... 

cat /etc/fstab 
# /etc/fstab: static file system information. 
# 
# <file system> <mount point>   <type>  <options>               <dump>  <pass> 
/dev/hda1       /               ext3    defaults,noatime,errors=remount-ro      0 0 
proc            /proc           proc    defaults                        0 0 
none            /dev/pts        devpts  gid=5,mode=20                   0 0 
/dev/hda2       swap            swap    defaults                        0 0 
/dev/hda3       /mnt            jfs     defaults,noatime                0 0



davy_gravy - member


OK. It is working again, but there is apparently some unexpected/undesirable behavior here. It is trying to mount the CDROM as XFS, and that is just not going to work. This seems to happen whenever an audio cd is in the tray at boot time.

I'm concerned that during conversion something was left in the wrong state.

mindbender, do we need to mention this on the XFS hda3 conversion article page?

===================

mindbender - Moderator

Quote

Quoting: davy_gravy


what did you do to fix it?

===================

davy_gravy - member

Simple to fix/workaround. I just make sure that there is no CD in the USB/CDROM drive when it boots. But why should a CD (an audio CD) in the drive trigger an attempt to mount something as XFS, when there is no XFS system on the hard drive anymore.

Not simple to figure out why it happend. Very unexpected that it happened. It *is* odd.

Has lb_worm seen anything like this? Anyone else who is using XFS or JFS, or has used convertfs seen this type of behavior?

I suspect that one of my apps (Mediaripper or one of its dependencies) may be an unwitting accomplice in this...

Do you think if I had a line in the fstab for the USB/CDburner it might keep it from happening? But, that would only make it try to mount the CD to the specified mount point...?

Retrieved from "http://buffalo.nas-central.org/w/index.php?title=Talk:Using_XFS_instead_of_ext3_(network_performance_boost)&oldid=8600"