On Fri, Sep 26, 2008 at 9:35 AM, John Drescher <drescherjm AT gmail DOT com>
wrote:
> On Fri, Sep 26, 2008 at 9:50 AM, Chris Picton <chris AT ecntelecoms DOT com>
> wrote:
>> On Fri, 2008-09-26 at 09:43 -0400, John Drescher wrote:
>>> > Just a quick test I ran using
>>> > dd if=/dev/zero of=/mnt/store/00/bigfile bs=1M count=8192
>>> > just to get a basic idea of speed:
>>> >
>>> > Raid 5: 8589934592 bytes (8.6 GB) copied, 302.948 seconds, 28.4 MB/s
>>> > Raid 0: 8589934592 bytes (8.6 GB) copied, 27.2641 seconds, 315 MB/s
>>> >
>>> > The Raid 5 will not be able to keep up with my network cards...
>>> >
>>> Something is very wrong with your setup. Are you using ext3?
>>
>> Yes - ext3
>>
>> I had not changed any of the raid stripe size or sys parameters - I just
>> wanted to get an idea of the speeds.
>>
>>
> I can help you with this.
>
> Here is what I get on a 5 drive software raid5
>
> # dd if=/dev/zero of=/bigfile bs=1M count=8192
> 8192+0 records in
> 8192+0 records out
> 8589934592 bytes (8.6 GB) copied, 44.8402 s, 192 MB/s
>
> # cat /proc/mdstat
> Personalities : [linear] [raid0] [raid1] [raid6] [raid5] [raid4] [multipath]
> md0 : active raid1 sde1[4] sdd1[3] sdc1[2] sdb1[1] sda1[0]
> 256896 blocks [5/5] [UUUUU]
> md1 : active raid5 sde3[4] sdd3[3] sdc3[2] sdb3[1] sda3[0]
> 46909440 blocks level 5, 64k chunk, algorithm 2 [5/5] [UUUUU]
>
> # mount | grep /dev/md1
> /dev/md1 on / type xfs (rw,noatime,sunit=128,swidth=512,ikeep,noquota)
>
>
> And the tuning param that I use:
>
> echo 1024 > /sys/block/md1/md/stripe_cache_size
>
> The reason for this and why I asked you for the filesystem is that
> raid5 is very slow if you write less than 1 stripe size if the stripe
> is not cached. The reason for the slowness is that if the block is not
> cached the os will need to read the block then calculate parity then
> write. This process is very slow. So above I extend the stripe cache.
> One of the problems with using ext3 without tuning is that by default
> it writes too frequently in small chunks. If you keep that I would
> look into the commit mount param and extend this to 10 or more
> seconds.
I have similar speeds with my raid 5 (software raid 500gb x4). Is
there more documentation on how I could tune the ext3 to get me over
100mb/s?
Using lvm
/dev/mapper/server7-server7_home on /home type ext3 (rw)
cat /proc/mdstat
Personalities : [raid1] [raid6] [raid5] [raid4]
md5 : active raid5 sda2[0] sdd2[3] sdc2[2] sdb2[1]
1462211904 blocks level 5, 64k chunk, algorithm 2 [4/4] [UUUU]
What should it be changed to ? How safe is it to do it while I already
have 80% capacity filled?
Thanks,
Lucas
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
