| To: | Tim Shimmin <tes@xxxxxxx>, linux-xfs@xxxxxxxxxxx |
|---|---|
| Subject: | Re: Speeding up XFS |
| From: | lawalsh <xfs@xxxxxxxxx> |
| Date: | Fri, 14 May 2004 11:05:12 -0700 |
| In-reply-to: | <20040505151243.V3275@boing.melbourne.sgi.com> |
| References: | <20040426165155.GA19148@em.ca> <200404261726.i3QHQ8Aj298772@zhadum.americas.sgi.com> <20040505151243.V3275@boing.melbourne.sgi.com> |
| Sender: | linux-xfs-bounce@xxxxxxxxxxx |
| User-agent: | Mozilla Thunderbird 0.6 (Windows/20040502) |
|
Haven't tried 64 and 128 yet, but 256 made ugly ascii error messages on my screen (generic mount failed message). I'd increased the #buffs to 8 and the bufsize to 262144, and got failures on mounting. #buffs=8 worked though and that alone seemed to improve "rm" speed noticably. It's on my "todo"l list to try 128 and 64. My memory might be too fragmented, dunno -- computer's memory too...:-). This was under 2.6.5 kernel. I wouldn't put too much weight on the problem until I've done some more testing to narrow it down, but thought I'd mention it as a data point. System I tried it on had been up for 15 days... Total mem=1G with 128M in high mem. -l (musing: I wonder if we will need DOS-style disk defragmenters for memory for systems with long uptimes and large memory...) :-) Tim Shimmin wrote: Up to 256Kb log buffers on Linux should work now. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | [PATCH] Laptop mode control script support for XFS *_centisecs sysctl values., Bart Samwel |
|---|---|
| Next by Date: | [Bug 328] freeze on XFS recovery on sparc64 with linux kernel 2.4.26, bugzilla-daemon |
| Previous by Thread: | Re: Speeding up XFS, Tim Shimmin |
| Next by Thread: | Re: Speeding up XFS, Chris Wedgwood |
| Indexes: | [Date] [Thread] [Top] [All Lists] |