thanks a lot Mark.
sorry i made a mistake just now, the correct report for the squid process
should be as follow, the one i posted just now is for runcache process. for
squid process, it seems that VSZ is not significantly larger than RSS.
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
nobody 19729 48.9 32.4 1303636 1301388 ? S Apr03 10682:44 squid -NsY
the return from top is
top - 15:06:17 up 23 days, 10:18, 4 users, load average: 1.20, 1.13, 1.06
Tasks: 66 total, 2 running, 64 sleeping, 0 stopped, 0 zombie
Cpu(s): 34.7% user, 15.5% system, 0.0% nice, 49.8% idle
Mem: 4010972k total, 4002048k used, 8924k free, 16264k buffers
Swap: 530136k total, 0k used, 530136k free, 1835588k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
19729 nobody 18 0 1271m 1.2g 1436 R 62.1 32.4 10687:31 squid
7563 root 12 0 956 956 760 R 5.6 0.0 0:00.06 top
according to the results, is there any problem with the memory status? sorry
for so many questions..thanks a lot.
From: "Mark Elsen" <mark.elsen@xxxxxxxxx>
To: "K Pang" <time_breath@xxxxxxxxxxx>
CC: squid-users@xxxxxxxxxxxxxxx
Subject: Re: high response time in the case of high file
descriptor usage
Date: Tue, 18 Apr 2006 05:52:57 +0200
>
> Hi, using ps, the return is:
> USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
> root 19724 0.0 0.0 2276 1056 ? S
Apr03 0:00 /bin/sh
> /usr/loca
>
> is the memory status related to slow response?
> i've no idea which is the total process size, pls advise. thanks a lot.
>
>
The total process size is reported under the column VSZ; which in
your case is 2276.
The part in memory is reported under the column RSS, which is 1056.
This means you have a problem; total memory avail. can only handle half of
SQUID code, which will make the squid process swapping, hence it
will become slow.
You need more physical memory.
M.
