From: sajeev nv (sajeev20@rediffmail.com)
Date: Wed Apr 10 2002 - 07:56:08 EDT
Hi,
System Config:
H/W:
Sun 420R ( Single CPU )
RAM:2048 MB
OS: Solaris 2.8 (with recommended patches)
S/W:
Oracle 8i Enterprise Edition-64bit rel 8.1.6.0.0
Iplanet 4.1 Enterprise Web Server with SP9
Iplanet LDAP 4.1
Parameters set in /etc/system are:
shmmax=14294967295
shmmin=1
shmmni=100
shmseg=10
semmsl=100
semmns=200
semmni=100
semopm=100
semvmx=32767
set rlim_fd_max=5293
set rlim_fd_cur=1024
set pidmax=100000
set maxuprc=99995
Parameters set in magnus.conf (iplanet)
ListenQ 1024
RqThrottle 512
StackSize 131072
CGIWaitPid on
MinCGIStubs 10
MaxCGIStubs 40
CGIStubIdleTimeout 45
AsyncDNS on
MaxProcs 3
MinAcceptThreadPerSocket 3
MaxAcceptThreadPerSocket 6
Hope the above details are sufficient.
Problem Description:
After a period ranging from 14-96hrs. the web server starts
becoming slow in response to the world and within 2 hrs. time it
stops responding.
During the above period the system behaves normally i.e. console
logons or even remote telnet sessions and also browsing from the
console appear to be working fine at normal speed.
Corrective measure:
The Web server is stopped and restarted manually and the system
comes back to normal without restarting the OS.
I have some output files taken during the above mentioned problem
state. the files are:
1] TOP
last pid: 10906; load averages: 1.25, 1.23, 1.06 14:38:05
122 processes: 120 sleeping, 1 running, 1 on cpu
Memory: 2048M real, 1025M free, 563M swap in use, 6940M swap
free
PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU
COMMAND
719 nobody 999 30 0 183M 101M run 43:36 87.86%
ns-httpd
446 oracle 1 58 0 271M 248M sleep 113:07 2.93%
oracle
8970 nobody 343 59 0 132M 49M sleep 2:55 1.92%
ns-httpd
10906 root 1 30 0 1792K 1232K cpu 0:00 1.15%
top
8966 nobody 340 59 0 132M 50M sleep 3:11 0.72%
ns-httpd
5562 nsuser 11 59 0 46M 41M sleep 2:33 0.56%
ns-slapd
10844 root 1 58 0 2608K 1672K sleep 0:01 0.35%
top
7228 root 1 59 0 23M 24M sleep 5:43 0.33%
Xsun
10345 root 1 59 0 8168K 5840K sleep 0:01 0.21%
dtterm
450 oracle 1 58 0 270M 240M sleep 14:41 0.16%
oracle
394 root 12 49 0 18M 18M sleep 3:06 0.12%
mibiisa
712 oracle 1 59 0 12M 6544K sleep 3:31 0.02%
tnslsnr
8612 root 8 59 0 8912K 6984K sleep 0:07 0.01%
dtwm
432 oracle 1 58 0 269M 239M sleep 0:20 0.00%
oracle
444 oracle 1 58 0 269M 245M sleep 0:01 0.00%
oracle
430 oracle 1 58 0 269M 239M sleep 0:49 0.00%
oracle
428 oracle 1 59 0 269M 239M sleep 0:27 0.00%
oracle
224 root 10 50 0 2720K 2016K sleep 0:22 0.00%
nscd
452 oracle 1 58 0 269M 239M sleep 0:04 0.00%
oracle
434 oracle 1 58 0 269M 243M sleep 0:03 0.00%
oracle
8625 root 1 58 0 2032K 1448K sleep 0:01 0.00%
rpc.rstatd
186 daemon 4 0 0 2488K 1776K sleep 0:00 0.00%
statd
152 root 4 0 0 2784K 1328K sleep 0:00 0.00%
keyserv
183 root 1 0 0 1864K 1184K sleep 0:00 0.00%
lockd
739 nobody 1 10 0 2072K 1184K sleep 0:00 0.00%
Cgistub
9020 nobody 1 10 0 2072K 1184K sleep 0:00 0.00%
Cgistub
9035 nobody 1 10 0 2072K 1184K sleep 0:00 0.00%
Cgistub
9044 nobody 1 10 0 2080K 472K sleep 0:00 0.00%
Cgistub
In the above top o/p the first process ns-httpd takes 87.86%
kernel time. Where as CPU idle time,iowait is 0.0%
2] mpstat
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr
sys wt idl
2 10 0 0 478 372 313 46 0 2 0 1380 8
4 1 86
3] vmstat
procs memory page disk faults
cpu
r b w swap free re mf pi po fr de sr s0 s1 s6 -- in sy
cs us sy id
0 0 0 7266952 1216840 7 10 3 0 0 0 0 4 3 0 0 378 1381
313 8 4 88
4] iostat
tty sd0 sd1 sd6 st11
cpu
tin tout kps tps serv kps tps serv kps tps serv kps tps serv
us sy wt id
0 93 37 4 174 26 3 5 0 0 0 0 0 66
8 4 1 87
I need to know why the above problem occurs for no apparent reason
and what is the solution?
Also I need to know if there is any patch required for Oracle 8i
and Solaris 2.8?
Regards,
SAJ
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:24:11 EDT