Thanks in advance for any help with this problem:
We are currently running Tru64 Unix 4.0F on a DECserver 2100/500MP with 256MB of memory and a swap area of roughly 200MB. We previously had been running this machine as our web/ftp server with Digital Unix 3.2B We have temporarily moved our web pages to a SunServer where we are getting very quick response. When we try using the latest httpd version 1.5 to run these pages on our Tru64 Unix box we get horrible response times. Same network and the machine is almost completely idle. When we were running 3.2B we had great response time on this Alpha machine. The response is almost 100 times worse than the SunServer, or the Alpha when we had 3.2B running. The former administrator of this machine no longer works for us and the notes for set-up were non-existent. 4.0F was installed with a Standard default choice.
Here is a snapshot using a utility when a single person hits the web page:
load averages: 0.01, 0.02, 0.00 01:23:22
45 processes: 1 running, 22 sleeping, 18 idle, 4 zombie
CPU states: 0.5% user, 0.0% nice, 1.5% system, 97.9% idle
Memory: Real: 36M/245M act/tot Virtual: 196M use/tot Free: 184M
PID USERNAME PRI NICE SIZE RES STATE TIME CPU COMMAND
609 WWW 44 0 3120K 581K sleep 0:00 0.40% <httpd>
617 root 44 0 2592K 393K run 0:03 0.30% <top>
610 WWW 44 0 3120K 532K sleep 0:00 0.20% <httpd>
608 WWW 44 0 3120K 573K sleep 0:00 0.10% <httpd>
561 root 44 0 7000K 1851K sleep 0:03 0.00% <Xdec>
467 root 42 0 6256K 2785K sleep 0:01 0.00% <insightd>
423 root 44 0 2848K 630K sleep 0:00 0.00% <cpq_mibs>
179 root 44 0 2280K 589K sleep 0:00 0.00% <ogated>
613 WWW 44 0 3120K 581K sleep 0:00 0.00% <httpd>
607 WWW 44 0 3120K 581K sleep 0:00 0.00% <httpd>
606 WWW 44 0 3120K 581K sleep 0:00 0.00% <httpd>
615 WWW 44 0 3120K 573K sleep 0:00 0.00% <httpd>
614 WWW 44 0 3120K 573K sleep 0:00 0.00% <httpd>
611 WWW 44 0 3120K 573K sleep 0:00 0.00% <httpd>
612 WWW 44 0 3120K 532K sleep 0:00 0.00% <httpd>
Here is a vmstat output:
(root) >vmstat -P
Total Physical Memory = 256.00 M
= 32768 pages
Physical Memory Clusters:
start_pfn end_pfn type size_pages / size_bytes
0 256 pal 256 / 2.00M
256 32759 os 32503 / 253.93M
32759 32768 pal 9 / 72.00k
Physical Memory Use:
start_pfn end_pfn type size_pages / size_bytes
256 280 unixtable 24 / 192.00k
280 283 scavenge 3 / 24.00k
283 614 text 331 / 2.59M
614 678 data 64 / 512.00k
678 790 bss 112 / 896.00k
790 794 cfgmgmt 4 / 32.00k
794 833 unixtable 39 / 312.00k
833 846 pmap 13 / 104.00k
846 1353 vmtables 507 / 3.96M
1353 32759 managed 31406 / 245.36M
============================
Total Physical Memory Use: 32503 / 253.93M
Managed Pages Break Down:
free pages = 20926
active pages = 1464
inactive pages = 2035
wired pages = 2528
ubc pages = 4458
==================
Total = 31411
WIRED Pages Break Down:
vm wired pages = 361
ubc wired pages = 0
meta data pages = 975
malloc pages = 630
contig pages = 74
user ptepages = 450
kernel ptepages = 27
free ptepages = 9
==================
Total = 2526
Any help, or suggestions (other than keeping the web pages on the Sun) would be most appreciated. Perhaps we are missing something with httpd.conf ? We are using the file as set up on our alpha when it was running Unix 3.2B.
I'd be glad to write up a summary, thanks again...
Received on Fri Feb 18 2000 - 05:50:19 NZDT