Messages in this thread | | | Date | Sat, 21 Feb 2004 18:46:49 -0800 | From | William Lee Irwin III <> | Subject | Re: 2.6.3-mm1 |
| |
On Tue, Feb 17, 2004 at 11:21:30PM -0800, Andrew Morton wrote: > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.3/2.6.3-mm1/ > - Added the dm-crypt driver: a crypto layer for device-mapper. > - Added the fbdev cursor API patch. Not sure what this does apart from > - There's a patch here to consolidate the 32->64 compat code for the IPC > syscalls. Needs testing on various 64-bit machines. > - Various random fixes to things.
Everything's happy here using things "in production" so to speak. I couldn't think of much else to be more informative/useful, so I just ran dbench 16 on 6 spindles simultaneously. cfq io scheduler, ext3 fs's, 768MB/600MHz Athlon, 5 47GB 10Krpm U160 disks, 1 73GB 10Krpm U160 disk, Adaptec 39160 dual channel U160 HBA. No idea if these numbers are good, bad, or what, but things certainly seem to be doing well. There's also some heavy NFS usage which isn't really performance-oriented things are doing well with also (cross-mounted /home/ and a number of nfsroot fs's). The earlier NFS problem report is being blamed on hardware (a potentially dead PMAD-AA ethernet controller in a distinct client machine, not anything in this machine) for now.
vmstat log attached.
==> /local/dbench/dbench.log.6 <== 0 62477 10.82 MB/sec Throughput 10.8175 MB/sec 16 procs
==> /mnt/d/dbench/dbench.log.6 <== 0 62477 9.85 MB/secc Throughput 9.84865 MB/sec 16 procs
==> /mnt/e/dbench/dbench.log.6 <== 0 62477 9.58 MB/secc Throughput 9.57926 MB/sec 16 procs
==> /mnt/f/dbench/dbench.log.6 <== 0 62477 9.62 MB/sec Throughput 9.61646 MB/sec 16 procs
==> /mnt/h/dbench/dbench.log.6 <== 0 62477 9.64 MB/sec Throughput 9.63958 MB/sec 16 procs
==> /share/mp3/dbench/dbench.log.6 <== 0 62477 9.62 MB/sec Throughput 9.62277 MB/sec 16 procs
-- wli
| |