unix sysadmin archives
Donation will make us pay more time on the project:
          

Thursday 20 October 2011

PICL bug causes Solaris 10 prtdiag to hang

The Solaris PICL framework provides information about the system configuration which it maintains in the PICL tree. I have an experience wherein Solaris 10 prtdiag is hanging. In order to fix this stop and start picld.

# top
load averages: 1582.95, 1462.52, 1345.91 22:57:54
8548 processes:8532 sleeping, 1 running, 1 zombie, 14 on cpu
CPU states: % idle, % user, % kernel, % iowait, % swap
Memory: 8064M real, 2747M free, 4123M swap in use, 9005M swap free
PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
13622 root 999 59 0 318M 297M sleep 370.3H 82.48% java
26222 root 1 0 0 0K 0K cpu/6 7:52 5.88% ps
25217 root 1 20 0 0K 0K sleep 11:59 5.49% ps
27618 root 1 0 0 0K 0K cpu/5 1:08 5.43% ps
27101 root 1 0 0 0K 0K cpu/4 3:31 5.16% ps

***You can see here that PID 13622 is using alot of CPU.
***And when you check this, it points to prtdiag 

# ps -ef | grep 13622

root 23066 13622 0 Oct 15 ? 0:00 /usr/bin/ctrun -l child -o pgrponly /bin/sh -c /usr/sbin/prtdiag
root 802 13622 0 Oct 15 ? 0:00 /usr/bin/ctrun -l child -o pgrponly /bin/sh -c /usr/sbin/prtdiag
root 28092 13622 0 Oct 15 ? 0:00 /usr/bin/ctrun -l child -o pgrponly /bin/sh -c /usr/sbin/prtdiag

***Restart the PICL
# svcadm restart picl

***Check the load via uptime

# uptime
1:26am up 50 day(s), 11:22, 3 users, load average: 1886.79, 1513.28, 1402.57

***After a couple of minutes check it again
# uptime
1:26am up 50 day(s), 11:23, 3 users, load average: 962.59, 1327.11, 1343.05

***You can observe a dramatic drop on the load
# top
load averages: 3.71, 367.87, 875.60 01:33:23
76 processes: 75 sleeping, 1 on cpu
CPU states: % idle, % user, % kernel, % iowait, % swap
Memory: 8064M real, 5630M free, 1103M swap in use, 12G swap free

PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
15726 root 1 42 0 108M 33M sleep 21:44 2.89% bptm
11116 root 1 52 0 108M 33M sleep 10:50 0.76% bptm
13622 root 78 59 0 215M 200M sleep 403.3H 0.12% java
15611 root 1 59 0 108M 67M sleep 5:09 0.10% bptm
16953 root 1 0 0 4432K 2160K cpu/11 0:00 0.08% top

No comments:

Post a Comment