RE: [SQU] squid tuning... again...

From: Wisnu Wijayanta <wisnu@dont-contact.us>
Date: Wed, 16 Aug 2000 10:18:24 +0700

Below, i've inserted the 5 minutes avg counters. Actually the cpu usage,
now... is fairly acceptable (30-40 percent). But, why the http req/min
starts to falling down slowly, is this normal behaviour on a busy day? By
means of our main link become more saturated or there are something else
goes wrong?

Before i ask you more... :), one of my responsibility is to design the
caching infrastructure for my corporation (with so many branches). Actuall
i've setting up about 5 to 10 production cache server, from MSP 2.0,
Netscape Proxy 3.52, Compaq TaskSmart and Squid. Some of this servers were
accessing the internet independently, some of them collaborate in a
parent-sibling models. And Forward i've looking for cache farm solution,
with Squid. Where can I found the reference to build this infrastructures?

Great thanks,

5 min avg count:

Cache Manager menu

----------------------------------------------------------------------------

----
sample_start_time = 966395710.792542 (Wed, 16 Aug 2000 03:15:10 GMT)
sample_end_time = 966396010.816034 (Wed, 16 Aug 2000 03:20:10 GMT)
client_http.requests = 28.904403/sec
client_http.hits = 3.686378/sec
client_http.errors = 0.000000/sec
client_http.kbytes_in = 11.845739/sec
client_http.kbytes_out = 76.960640/sec
client_http.all_median_svc_time = 0.030657 seconds
client_http.miss_median_svc_time = 2.020665 seconds
client_http.nm_median_svc_time = 0.008653 seconds
client_http.nh_median_svc_time = 2.376078 seconds
client_http.hit_median_svc_time = 0.012346 seconds
server.all.requests = 11.609091/sec
server.all.errors = 0.000000/sec
server.all.kbytes_in = 55.192345/sec
server.all.kbytes_out = 6.776136/sec
server.http.requests = 11.539096/sec
server.http.errors = 0.000000/sec
server.http.kbytes_in = 55.119017/sec
server.http.kbytes_out = 6.732806/sec
server.ftp.requests = 0.000000/sec
server.ftp.errors = 0.000000/sec
server.ftp.kbytes_in = 0.000000/sec
server.ftp.kbytes_out = 0.000000/sec
server.other.requests = 0.069995/sec
server.other.errors = 0.000000/sec
server.other.kbytes_in = 0.069995/sec
server.other.kbytes_out = 0.043330/sec
icp.pkts_sent = 1.183241/sec
icp.pkts_recv = 1.183241/sec
icp.queries_sent = 0.000000/sec
icp.replies_sent = 1.183241/sec
icp.queries_recv = 1.183241/sec
icp.replies_recv = 0.000000/sec
icp.replies_queued = 0.000000/sec
icp.query_timeouts = 0.000000/sec
icp.kbytes_sent = 0.086660/sec
icp.kbytes_recv = 0.089993/sec
icp.q_kbytes_sent = 0.000000/sec
icp.r_kbytes_sent = 0.086660/sec
icp.q_kbytes_recv = 0.089993/sec
icp.r_kbytes_recv = 0.000000/sec
icp.query_median_svc_time = 0.000000 seconds
icp.reply_median_svc_time = 0.000000 seconds
dns.median_svc_time = 0.923949 seconds
unlink.requests = 0.399969/sec
page_faults = 1.773194/sec
select_loops = 138.945786/sec
select_fds = 177.629424/sec
average_select_fd_period = 0.005590/fd
median_select_fds = 0.000000
swap_files_cleaned = 0.000000/sec
aborted_requests = 4.086347/sec
syscalls.polls = 479.179144/sec
syscalls.disk.opens = 3.883029/sec
syscalls.disk.closes = 3.879696/sec
syscalls.disk.reads = 3.009764/sec
syscalls.disk.writes = 40.973458/sec
syscalls.disk.seeks = 0.083327/sec
syscalls.disk.unlinks = 0.000000/sec
syscalls.sock.accepts = 44.153209/sec
syscalls.sock.sockets = 9.865894/sec
syscalls.sock.connects = 9.412596/sec
syscalls.sock.binds = 0.000000/sec
syscalls.sock.closes = 31.527531/sec
syscalls.sock.reads = 93.529343/sec
syscalls.sock.writes = 82.303555/sec
syscalls.sock.recvfroms = 3.566387/sec
syscalls.sock.sendtos = 1.806525/sec
cpu_time = 157.250000 seconds
wall_time = 300.023492 seconds
cpu_usage = 52.412562%
----------------------------------------------------------------------------
----
Generated Wed, 16 Aug 2000 03:20:28 GMT, by
cachemgr.cgi/2.3.STABLE4@kilimanjaro.ai.astra.co.id
I'm running out of ideas.  Is your CPU usage still high?  Maybe
send the output of the "5 Minute Average of Counters" from
the cachemgr.
--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Tue Aug 15 2000 - 21:25:06 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:52 MST