lkml.org 
[lkml]   [2005]   [Jun]   [14]   [last100]   RSS Feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: Tuning ext3 for large disk arrays
FromBadari Pulavarty <>
Date14 Jun 2005 17:59:20 -0700
On Tue, 2005-06-14 at 18:13, Peter Chubb wrote:
> >>>>> "Badari" == Badari Pulavarty <pbadari@us.ibm.com> writes:
> 
> Badari> What kernel are you running ?  Does the kernel has ext3
> Badari> "reservation" support enabled ? 
> 2.6.11, and yes, although it's not set in the mount options.

2.6.11 has "reservation" turned ON by default. You should be
fine.

> 
> 
> Badari> Do you see performance problem
> Badari> with "read" tests also ? 
> 
> Yes.

Hmm. 

> 
> Badari> And also, does the write test writes
> Badari> to multiple files in the same directory ? Or multiple threads
> Badari> writing to same file ?
> 
> It's standard iozone --- mutliple processes writing to multiple files in
> the same directory.

If you are running multiple files in the same directory, without
reservations files will be badly fragmented. 

Can you run "filefrag" on the files ?

(I am also assuming each process writes to its own file).


Thanks,
Badari

> 
> All the mount options are the defaults.
> 
> A sample test is:
> 
> mdadm -Ss
> mdadm -C /dev/md0 -l 0 -c 1024 -n 2 -R  /dev/sdc /dev/sdk 
> mke2fs -j  /dev/md0 
> mount /dev/md0 /shift/oplapro97/data01
> cd /shift/oplapro97/data01
> iozone -MCew -t1 -f IOZONE -s32g -r256k -i0 -i1 > ~/tests/unsw_ext3/iozone_2_disk_1stream_ext3
> rm -rf /shift/oplapro97/data01/*
> iozone -MCew -t2 -f IOZONE -s32g -r256k -i0 -i1 > ~/tests/unsw_ext3/iozone_2_disk_2stream_ext3
> 

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-06-15 03:26    [from the cache]
©2003-2008