| To: | linux-xfs <linux-xfs@xxxxxxxxxxx> |
|---|---|
| Subject: | Re: Defrag pitfals (was Re: Support for XFS File systems) |
| From: | Florin Andrei <florin@xxxxxxx> |
| Date: | 13 Feb 2003 10:40:00 -0800 |
| In-reply-to: | <Pine.LNX.4.44.0302121952080.7758-100000@kenzo.iwr.uni-heidelberg.de> |
| References: | <Pine.LNX.4.44.0302121952080.7758-100000@kenzo.iwr.uni-heidelberg.de> |
| Reply-to: | linux-xfs@xxxxxxxxxxx |
| Sender: | linux-xfs-bounce@xxxxxxxxxxx |
On Wed, 2003-02-12 at 11:28, Bogdan Costescu wrote: > After switching to XFS and (mea culpa!) forgetting to set up xfs_fsr to be > run by cron, the read speed would be similar after several days-weeks. > However, by using xfs_fsr we go back to reading with around 20 MiB/s even > for a pretty full FS. There were many helpful answers posted to the list, surely you can pick a few good ones and try. A generic method that seems to work pretty much everywhere and provides good results is to avoid the disk usage to grow too much. A filesystem that's at 95% usage will usually be quite fragmented. Stay away from that. By doing some tests you may be able to determine which is - in your case - the disk usage limit which triggers the fragmentation problems. -- Florin Andrei A cloned dog doesn't know the old tricks. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | TAKE 880764 - fix another transaction ordering problem, Steve Lord |
|---|---|
| Next by Date: | TAKE - fix dirty buffer leak (in error path), Christoph Hellwig |
| Previous by Thread: | Re: Defrag pitfals (was Re: Support for XFS File systems), Bogdan Costescu |
| Next by Thread: | RE: Support for XFS File systems, Marc Kaplan |
| Indexes: | [Date] [Thread] [Top] [All Lists] |