Hi List,
I use a box with 4 harddisks with Raid5 and XFS. After writing a File the
File appears to change when
reading it. I've done the following to demonstrate the issue:
Created 2 files 100 MB each with /dev/zero as source and compared the
files 3
times.
[root@mss archive]# dd if=/dev/zero of=testfile1 bs=1M count=100
100+0 Records in
100+0 Records out
[root@mss archive]# dd if=/dev/zero of=testfile2 bs=1M count=100
100+0 Records in
100+0 Records out
[root@mss archive]# compare testfile1 testfile2 -a
14614544 (0xdf0010) 0x7f != 0x00 ^? ^@
14614545 (0xdf0011) 0xff != 0x00 ~^? ^@
71893008 (0x4490010) 0x00 != 0x7f ^@ ^?
71893009 (0x4490011) 0x00 != 0xff ^@ ~^?
[root@mss archive]# compare testfile1 testfile2 -a
44630032 (0x2a90010) 0x7f != 0x00 ^? ^@
44630033 (0x2a90011) 0xff != 0x00 ~^? ^@
58785808 (0x3810010) 0x7f != 0x00 ^? ^@
58785809 (0x3810011) 0xff != 0x00 ~^? ^@
[root@mss archive]# compare testfile1 testfile2 -a
8716304 (0x850010) 0x7f != 0x00 ^? ^@
8716305 (0x850011) 0xff != 0x00 ~^? ^@
31260688 (0x1dd0010) 0x7f != 0x00 ^? ^@
31260689 (0x1dd0011) 0xff != 0x00 ~^? ^@
31326224 (0x1de0010) 0x7f != 0x00 ^? ^@
31326225 (0x1de0011) 0xff != 0x00 ~^? ^@
85385232 (0x516e010) 0x7f != 0x00 ^? ^@
85385233 (0x516e011) 0xff != 0x00 ~^? ^@
104136720 (0x6350010) 0x00 != 0x7f ^@ ^?
104136721 (0x6350011) 0x00 != 0xff ^@ ~^?
you can see this yields different problems every compare on both sides ...
This looks to me like the data is written correctly to disk and the
problem
happens when reading the data.
The Kerne I use is 2.4.22-ac3. I recently updated to 2.4.23-rc1 but the
Problem is still present.
After some Tests and some mails to the XFS List I've read the Files with
'dd' directly from the raiddevice. The Problem is still there - When I put
the Files on another volume - even another raid5 array on the same box the
errors didn't change anymore .. like a snapshot of the current error.
It's a live system with some valuable data for me - the /dev/zero tests
weher just to clarify the problem.
Here some additional Informations from the box:
[root@mss /]# cat /etc/raidtab
raiddev /dev/md0
raid-level 5
chunk-size 64k
persistent-superblock 1
nr-raid-disks 4
device /dev/hde1
raid-disk 0
device /dev/hdg1
raid-disk 1
device /dev/hdi1
raid-disk 2
device /dev/hdk1
raid-disk 3
raiddev /dev/md1
raid-level 5
chunk-size 64k
persistent-superblock 1
nr-raid-disks 3
device /dev/hdf1
raid-disk 0
device /dev/hdh1
raid-disk 1
device /dev/hdj1
raid-disk 2
[root@mss /]# cat /proc/mdstat
Personalities : [linear] [raid0] [raid1] [raid5] [multipath]
read_ahead 1024 sectors
md0 : active raid5 hdk1[3] hdi1[2] hdg1[1] hde1[0]
360160896 blocks level 5, 64k chunk, algorithm 0 [4/4] [UUUU]
md1 : active raid5 hdj1[2] hdh1[1] hdf1[0]
156280064 blocks level 5, 64k chunk, algorithm 0 [3/3] [UUU]
unused devices: <none>
[root@mss /]# cat /proc/ide/hd[e-l]/model
Maxtor 4G120J6
SAMSUNG SV1604N
Maxtor 4G120J6
SAMSUNG SV1604N
Maxtor 4G120J6
Maxtor 4D080H4
Maxtor 4G120J6
Maxtor 98196H8
If you need anthing else or this is the wrong way to report an error,
please
inform me.
Thanks
Stefan Hagendorn
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
