Re: 1x 3ware controllers vs. 2x 3ware controllers
From: Mark Watts <hidden>
Date: 2004-07-30 10:16:42
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
quoted
With the performance issues I'm seeing with the 8506-4LP's I have, Iwouldn'tquoted
recommend them to anyone currently... The thread on LKML from a few days ago about mke2fs -j and 8506-4LP says it all, but basically and reasonable amount of I/O brings a Dual Opteronsystemquoted
to its knees. And by reasonable I mean copying ISO images from a usb2 drive to a raid 5(4 xquoted
250GB maxtor) or even just formatting a 600MB partition.I've been able to get nearly 1.5 Gbits/sec off of an 8506-8, with 8 250GB disks... of course those are all 1 megabyte reads (but from random locations). Is your problem related to # of IO's, or size of transfer?
# hdparm -tT /dev/sda /dev/sda: Timing buffer-cache reads: 2232 MB in 2.00 seconds = 1115.61 MB/sec Timing buffered disk reads: 190 MB in 3.03 seconds = 62.76 MB/sec bonnie++ concurs with this and gives us about 22MB/sec write performance. Basically anything involving I/O is causing the system to crawl. I can reliably reproduce this by doing: mke2fs -j /dev/sda7 Where sda7 is a 600GB partition. Basically the system load goes orbital and responsivness goes out the window. gkrellm only shows ~1MB/sec of disk activity and the processors (Dual Opteron) are twiddling their thumbs (top shows 100% 'wa' state for both processors), yet I can hardly do anything with any open applications. Kernel is a 2.6.8rc1 kernel.org job. - -- Mark Watts Senior Systems Engineer QinetiQ Trusted Information Management Trusted Solutions and Services group GPG Public Key ID: 455420ED -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFBCiAKBn4EFUVUIO0RAsmIAJ4zaXUPemotyUDFTjj9soa4ZSD+GwCfYQMg jxmu5wO9o4XXh+0p54zKVio= =zWAH -----END PGP SIGNATURE-----