i'm running into a problem where writing to a file will cause a 5 to 15 second pause, this occurs only on writes. So far i have remounted the filesystem with data=writeback as an option, and set /sys/block/sdb/queue/max_sectors_kb to 64 to shorten the queue.
Are there any other things i can try to solve this ?
Are you reading and writing a large number of files?
mount -o remount,noatime,nodiratime /mountpoint
Can also help if you don't need to know the file's last access time.
Which IO scheduler are you using? Deadline/Anticipatory have some odd corner cases that will cause issues. CFQ is generally pretty good, but, the others might be better suited depending on your application: http://www.linux-mag.com/id/7564/2/
Are barriers enabled or disabled? Some controllers support barriers, but write performance suffers. mount -o remount,nobarrier /mountpoint or possibly mount -o remount,barrier=0 /mountpoint
I would recommend upgrading to ext4, you can do this in-place and it does have better write performance.
If you're reading and writing to the FS at the same time, then mounting with data=journal will improve performance dramatically.
The next thing to try is to tweak the bdflush paramater - to write dirty buffers more often.
You can check the disks you're writing to, if its a RAID array there could be problems with the drives; alternatively make sure the write cache is on.
Is your file system filled up. When file system space usage goes near 100% many file systems (and ext3 is not the worst) become slower and slower.
