performance of the 2.1.131-ac11 kernels, when using the HJ's knfsd
utilities is pretty good.
For instance:
(a quick test, only 100 meg in size file..)
# ./bonnie.solaris -d /auto/pdsfdv01
File '/auto/pdsfdv01/Bonnie.16680', size: 104857600
Writing with putc()...done
Rewriting...done
Writing intelligently...done
Reading with getc()...done
Reading intelligently...done
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
-------Sequential Output-------- ---Sequential Input--
--Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block---
--Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU
/sec %CPU
pdsflx16 1024 1575 19.7 2809 5.5 2060 5.0 3056 41.5 4241 5.6
28.7 3.5
starsu00 1024 6532 65.4 5705 15.1 4044 12.7 3683 48.8 5334 9.3
94.0 5.1
pdsfdv01 1024 6822 67.8 5988 17.7 2734 11.1 2663 38.7 5001 10.8
89.1 6.2
starsu00 was writing to 2.1.125 w/knfsd patches.
pdsfdv01 was writing to 2.1.131-ac11 kernel, w/raid0145 patches, and
knfsd utilities.
pdsflx16 is a Linux 2.0.35, running user NFS.
The hardware is the same; all running Intel 400Mhz/256MB ram, 100baseT
Full duplex, fully switched network.
client was Solaris 2.6, 4 CPU, QFE channel bundled (400MB pipe).
but, the biggest improvement comes when you have lots of clients hitting
the same NFS server. unfsd would just slow to a crawl, where knfsd just
keeps going, and going, and going...
-- ------------------------+-------------------------------------------------- Thomas Davis | PDSF Project Leader tadavis@lbl.gov | (510) 486-4524 | "Only a petabyte of data this year?"- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/