Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi,
sorry again,
The case (1)-(3) in my previous mail has no LVM.
put it all together,
1. DRBD 8.3.10 + LVM2(RHEL 5.2)
We are trying to configure a DRBD resource as a Physical Volume,
using examples from the following document.
http://www.drbd.org/users-guide/s-lvm-drbd-as-pv.html
There is a performance problem.
# iostat -xmd 10 /dev/cciss/c0d3
-------------------------------------------
| wMB/s | w/s
-------------------------------------------
hdd | 68.6 | 805
drbd | 47.3 | 910
lvm on drbd | 31.8 | 8140
-------------------------------------------
It seems that LVM2 included in RHEL 5.2 is relatively old,
so we have to move into its latest version.
Which LVM version is better suited for DRBD 8.3.10?
2. DRBD 8.3.10 + RHEL 6.1
with ext4 or xfs, DRBD device gives errors.
We don't use LVM in this case.
- ext4(Read-only file system )
# dd if=/dev/zero of=/mnt/test.dat bs=1M count=10000 conv=notrunc,fsync
dd: writing `/mnt/test.dat': Read-only file system
7385+0 records in
7384+0 records out
7743180800 bytes (7.7 GB) copied, 37.1128 s, 209 MB/s
-xfs(Input/output error )
# dd if=/dev/zero of=/mnt/test.dat bs=1M count=10000 conv=notrunc,fsync
dd: writing `/mnt/test.dat': Input/output error
4715+0 records in
4714+0 records out
4943020032 bytes (4.9 GB) copied, 12.7228 s, 389 MB/s
-dev/drbd0 directly (No error)
# dd if=/dev/zero of=/dev/drbd0 bs=1M count=10000 conv=notrunc,fsync
10000+0 records in
10000+0 records out
10485760000 bytes (10 GB) copied, 93.3198 s, 112 MB/s
Thanks,
Junko IKEDA
2011年5月24日14:00 Junko IKEDA <tsukishima.ha at gmail.com>:
> Hi,
>
>> "trapped in I/O error" is supposed to mean what, in this context?
>> Can you elaborate or rephrase?
>
> Here is the testing environment and logs;
>
> OS
> Red Hat Enterprise Linux Server release 6.1 (Santiago)
>
> kernel
> Linux node1 2.6.32-131.0.15.el6.x86_64 #1 SMP Tue May 10 15:42:40 EDT
> 2011 x86_64 x86_64 x86_64 GNU/Linux
>
> Hardware
> HP DL2000 (DL170e x 2）
>
> HDD
> SmartArray P410 SAS SSD 60GB x 11(RAID5)
>
> (1) ext4
> first, DRBD works well.
>
> # cat /proc/drbd
> version: 8.3.10 (api:88/proto:86-96)
> GIT-hash: 5c0b0469666682443d4785d90a2c603378f9017b build by
> root at node1, 2011-05-21 00:02:00
> 0: cs:Connected ro:Secondary/Primary ds:UpToDate/UpToDate C r-----
> ns:0 nr:11922760 dw:11922760 dr:0 al:0 bm:169 lo:0 pe:0 ua:0 ap:0
> ep:1 wo:b oos:0
>
> after mounting ext4 filesystem, run dd like this;
>
> # dd if=/dev/zero of=/mnt/test.dat bs=1M count=10000 conv=notrunc,fsync
> dd: writing `/mnt/test.dat': Read-only file system
> 7385+0 records in
> 7384+0 records out
> 7743180800 bytes (7.7 GB) copied, 37.1128 s, 209 MB/s
>
> For some reason dd gives an error.
> /proc/mounts shows the filesystem status as "ro", but mount command shows "rw".
>
> # cat /proc/mounts
> ...
> /dev/drbd0 /mnt ext4 ro,seclabel,relatime,barrier=1,data=ordered 0 0
>
> # mount
> ...
> /dev/drbd0 on /mnt type ext4 (rw)
>
> # tail -f /var/log/messages
> May 24 10:53:40 node1 kernel: EXT4-fs (drbd0): mounted filesystem with
> ordered data mode
> May 24 10:54:11 node1 kernel: JBD2: I/O error detected when updating
> journal superblock for drbd0-8.
> May 24 10:54:44 node1 kernel: Aborting journal on device drbd0-8.
> May 24 10:54:44 node1 kernel: JBD2: I/O error detected when updating
> journal superblock for drbd0-8.
> May 24 10:54:44 node1 kernel: EXT4-fs error (device drbd0):
> ext4_journal_start_sb: Detected aborted journal
> May 24 10:54:44 node1 kernel: EXT4-fs (drbd0): Remounting filesystem read-only
> May 24 10:54:44 node1 kernel: EXT4-fs error (device drbd0) in
> ext4_da_writepages: IO failure
> May 24 10:54:44 node1 kernel: EXT4-fs (drbd0): ext4_da_writepages:
> jbd2_start: 12288 pages, ino 12; err -30
> May 24 10:54:44 node1 kernel:
>
> (2) xfs
> after mounting xfs filesystem, run dd in the same way as case (1).
>
> # dd if=/dev/zero of=/mnt/test.dat bs=1M count=10000 conv=notrunc,fsync
> dd: writing `/mnt/test.dat': Input/output error
> 4715+0 records in
> 4714+0 records out
> 4943020032 bytes (4.9 GB) copied, 12.7228 s, 389 MB/s
>
> In this case, dd gives "Input/output error".
>
> syslog said;
> May 24 11:08:32 node1 kernel: Filesystem "drbd0": Disabling barriers,
> trial barrier write failed
> May 24 11:08:32 node1 kernel: XFS mounting filesystem drbd0
> May 24 11:09:02 node1 kernel: I/O error in filesystem ("drbd0")
> meta-data dev drbd0 block 0x2711172 ("xlog_iodone") error 95 buf
> count 3072
> May 24 11:09:02 node1 kernel: xfs_force_shutdown(drbd0,0x2) called
> from line 904 of file fs/xfs/xfs_log.c. Return address =
> 0xffffffffa038a1db
> May 24 11:09:02 node1 kernel: Filesystem "drbd0": Log I/O Error
> Detected. Shutting down filesystem: drbd0
> May 24 11:09:02 node1 kernel: Please umount the filesystem, and
> rectify the problem(s)
> May 24 11:09:32 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
> May 24 11:10:02 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
> May 24 11:10:32 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
> May 24 11:11:02 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
>
> after unmounting;
> May 24 11:12:57 node1 kernel: xfs_force_shutdown(drbd0,0x1) called
> from line 1062 of file fs/xfs/linux-2.6/xfs_buf.c. Return address =
> 0xffffffffa03a6239
> May 24 11:12:57 node1 kernel: xfs_force_shutdown(drbd0,0x1) called
> from line 1062 of file fs/xfs/linux-2.6/xfs_buf.c. Return address =
> 0xffffffffa03a6239
> May 24 11:12:58 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
> May 24 11:12:58 node1 kernel: Filesystem "drbd0": xfs_log_force: error
> 5 returned.
>
> (3) dev/drbd0
> dd dev/drbd0 directly.
>
> # dd if=/dev/zero of=/dev/drbd0 bs=1M count=10000 conv=notrunc,fsync
> 10000+0 records in
> 10000+0 records out
> 10485760000 bytes (10 GB) copied, 93.3198 s, 112 MB/s
>
> No error.
>
> Thanks,
> Junko IKEDA
>