Need your help about Memcache!
Jeffrey Ng
jeffreyn at gmail.com
Mon Feb 6 10:04:45 UTC 2006
04:45:06 up 54 days, 13:44, 3 users, load average: 0.45, 0.38, 0.53
78 processes: 77 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: cpu user nice system irq softirq iowait idle
total 10.0% 0.0% 3.6% 0.4% 1.6% 7.6% 376.0%
cpu00 7.5% 0.0% 2.6% 0.4% 1.9% 2.4% 84.8%
cpu01 0.0% 0.0% 0.2% 0.0% 0.0% 0.2% 99.5%
cpu02 1.9% 0.0% 0.4% 0.0% 0.0% 2.9% 94.6%
cpu03 0.2% 0.0% 0.0% 0.0% 0.0% 2.2% 97.5%
Mem: 3113604k av, 2726528k used, 387076k free, 0k shrd, 199908k
buff
1832240k active, 533308k inactive
Swap: 4192956k av, 521156k used, 3671800k free 771504k
cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
4523 memcache 16 0 243M 242M 432 S 3.9 7.9 1:54 1 memcached
6881 root 16 0 36828 23M 5296 S 0.2 0.7 0:20 0 httpd
6815 root 16 0 1768 692 524 S 0.0 0.0 0:00 2 syslogd
6840 root 16 0 3272 1384 1160 S 0.0 0.0 0:01 2 sshd
6854 qmails 16 0 1616 360 292 S 0.0 0.0 0:01 0 qmail-send
6855 qmaild 16 0 1640 492 440 S 0.0 0.0 0:00 1 tcpserver
6857 qmaill 16 0 1584 424 368 S 0.0 0.0 0:00 1 splogger
6858 root 18 0 1584 300 244 S 0.0 0.0 0:00 2
qmail-lspawn
6859 qmailr 16 0 1584 332 276 S 0.0 0.0 0:00 2
qmail-rspawn
6860 qmailq 16 0 1576 316 260 S 0.0 0.0 0:00 3
qmail-clean
6890 root 16 0 1632 600 528 S 0.0 0.0 0:00 0 crond
here's our top results. what other information will be helpful for you to
help us?
Jeffrey
On 2/6/06, Marcel Holan <mh at petamem.com> wrote:
>
> On Mon, Feb 06, 2006 at 05:11:06PM +0800, Jeffrey Ng wrote:
> > I am the coworker of Swansun. We are currently running apache and
> memcache
> > on the same machine. Maybe sometimes apache would use more RAM than it
> > should. Will that cause the problem swansun talked about? And how should
> we
> > prevent that kind of problem when we have apache and memcache on the
> same
> > VPS?
>
> Without some more quantitative information, help is difficult. Of course,
> the
> machine has to have enough ram for both apache as well as memcached and
> evtl. other processes running there. by forcing it to swap, all benefits
> from
> a fast RAM-retrieval (of precomputed data) are gone.
>
> Never forget, there's almost always a space/speed tradeoff.
>
> --
> best regards
> Marcel Holan
>
> project manager R&D
>
> +----------------------------------------------------------------------------+
> PetaMem s.r.o., Ocelarska 1, 190 00 Praha, Czech Republic -
> www.petamem.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.danga.com/pipermail/memcached/attachments/20060206/f99a54ed/attachment.htm
More information about the memcached
mailing list