Hi Sebastian,
I don't know if this will help at all, but I remember having a
Cyrus-IMAP server that always had a fair amount of disk I/O and tons of
open files that would exhibit the similar behavior. Unfortunately, it
wasn't running GENERIC (had RAIDFRAME and a couple of other things
enabled), it was running an old OpenBSD 3.6 (amd64) system and I didn't
have time to debug the issue since it was a production box, so I just
changed the /usr/src/sys/uvm/uvm_map.h KMAP_ENT #define to be 4000
instead of 1000, recompiled the kernel, and also added crontab entries
to stop and start most of the daemons running on that machine
frequently, and that kept it up long enough to migrate the data off of it.I know it's incredibly kludgey and hackish, but it worked temporarily
Best of luck,
TicoSebastian Reitenbach wrote:
>Hi list,
| Linus Torvalds | Linux 2.6.21 |
| Greg Kroah-Hartman | [PATCH 002/196] Chinese: rephrase English introduction in HOWTO |
| Con Kolivas | Re: [PATCH][RSDL-mm 0/7] RSDL cpu scheduler for 2.6.21-rc3-mm2 |
| Andrew Morton | echo mem > /sys/power/state |
git: | |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
| Gerrit Renker | [PATCH 27/37] dccp: Integration of dynamic feature activation - part 2 (server side) |
| David Miller | Re: [GIT]: Networking |
| Michael S. Tsirkin | Re: [RFC PATCH v2 03/19] vbus: add connection-client helper infrastructure |
