View previous topic :: View next topic |
Author |
Message |
TheForumTroll n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/122465441450da3fdae85c0.png)
Joined: 17 Jan 2006 Posts: 8
|
Posted: Sun Dec 23, 2012 2:01 am Post subject: Measure the impact of different commit intervals (ext4) |
|
|
I'm trying to figure out a way to measure the differences in performance when using different commit intervals on a file system (ext4 in this case). Any input on how to achieve this would be appreciated.
I tried searching but only came up with advice on what interval to use (but no reason as to why and when) and some bug discussions. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
TheForumTroll n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/122465441450da3fdae85c0.png)
Joined: 17 Jan 2006 Posts: 8
|
Posted: Wed Dec 26, 2012 12:03 am Post subject: |
|
|
Talking to myself here, but I tried using the disk tests in the Phoronix test suite. Didn't see any difference at the default(5) commit and at 10 minutes (600) so my guess is that they call fsync. I seem to be in way over my head here, so I'm still hoping someone can give me a hint. According to Theodore Ts'o a high interval should have a big impact on performance. But how to test it... ![Rolling Eyes :roll:](images/smiles/icon_rolleyes.gif) |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
Ant P. Watchman
![Watchman Watchman](/images/ranks/rank-G-2-watchman.gif)
Joined: 18 Apr 2009 Posts: 6920
|
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
TheForumTroll n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/122465441450da3fdae85c0.png)
Joined: 17 Jan 2006 Posts: 8
|
Posted: Wed Dec 26, 2012 1:00 am Post subject: |
|
|
Looks like it might. Thank you ![Very Happy :D](images/smiles/icon_biggrin.gif) |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
|