I have a server that is mail/web/shell for several thousand users. I have been watching the I/O load go up over time, and although it isn't causing any apparent issues yet, I didn't want to wait for a complaint. What is a good way to go about figuring out what is causing the I/O load to go up?
I started reading about the various block device schedulers (which seems to in many cases devolve into a religious argument). I decided to give "deadline" a try, and that has cut the I/O wait percentage just about in half (and smoothed it out as well). Are there any "gotchas" I should be on the lookout for? The primary storage is on an iSCSI SAN (Dell EqualLogic PS6000X), so I was also wondering what might happen if I set the OS scheduler to "noop" and let the SAN do all the scheduling (under the assumption that since it knows the actual hardware layout, it could do a better job). Has anybody tried that? -- Chris Adams <cmad...@hiwaay.net> Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. _______________________________________________ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
