Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

what can cause these kernel statistics (as reported by top) to be so high?

Kernel: 152661 ctxsw, 2475 trap, 99065 intr, 1449 syscall, 3 fork, 2373 flt

Usually, my system has much lower values, e.g.

Kernel: 487 ctxsw, 3 trap, 904 intr, 435 syscall, 3 flt

but every now and then the numbers go up and the OS freezes. The load is always <1.

Thank you!

Edit:

$ vmstat
 kthr      memory            page            disk          faults      cpu
 r b w   swap  free  re  mf pi po fr de sr s0 s1 s3 s4   in   sy   cs us sy id
 1 0 0 2806128 2818224 43 207 0 0  0  0  5  4  4  0 10 9954  510 3740  0  2 98

$ prstat
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
   658 root        0K    0K sleep   60    -   2:15:27 0.5% nfsd_kproc/157
   245 root        0K    0K sleep   99  -20   0:16:08 0.2% zpool-volume/166
   577 root        0K    0K sleep   60    -   0:00:09 0.0% lockd_kproc/24
  8195 root       11M 4788K cpu0    49    0   0:00:00 0.0% prstat/1
   617 root       53M   36M cpu3    59    0   0:00:34 0.0% fmd/29
   117 root     2144K 1288K sleep   59    0   0:00:00 0.0% pfexecd/3
   136 root       13M 4824K sleep   59    0   0:00:00 0.0% syseventd/19
    46 root       17M 8260K sleep   59    0   0:00:01 0.0% dlmgmtd/20
    42 netcfg   3892K 2900K sleep   59    0   0:00:00 0.0% netcfgd/4
    94 daemon     14M 4824K sleep   59    0   0:00:00 0.0% kcfd/3
   614 daemon     12M 2068K sleep   59    0   0:00:00 0.0% nfsmapid/3
   708 hpsmh      24M 6256K sleep   59    0   0:00:00 0.0% hpsmhd/1
    13 root       19M   18M sleep   59    0   0:00:14 0.0% svc.configd/18
    11 root       24M   14M sleep   59    0   0:00:04 0.0% svc.startd/16
    71 netadm   4272K 2908K sleep   59    0   0:00:00 0.0% ipmgmtd/5
Total: 78 processes, 930 lwps, load averages: 0.40, 0.44, 0.46

Edit 2: stats just before the crash:

$ vmstat
 kthr      memory            page            disk          faults      cpu
 r b w   swap  free  re  mf pi po fr de sr s0 s1 s3 s4   in   sy   cs us sy id
 1 0 0 2368992 2330108 41 216 0 0  0  0 130 4  4  0 64 39092 486 23076 0  7 93

$ prstat
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
   453 root        0K    0K sleep   99  -20   0:05:09 0.5% zpool-volume/166
   581 root        0K    0K sleep   60    -   0:20:36 0.4% nfsd_kproc/128
  1819 root       11M 6036K sleep   49    0   0:00:01 0.0% bash/1
   548 root        0K    0K sleep   60    -   0:00:16 0.0% lockd_kproc/12
     5 root        0K    0K sleep   99  -20   0:00:11 0.0% zpool-rpool/166
  1818 root       18M 5392K sleep   59    0   0:00:00 0.0% sshd/1
   555 root       58M   42M sleep   59    0   0:00:25 0.0% fmd/29
  3528 root       11M 5092K cpu5    59    0   0:00:00 0.0% prstat/1
     6 root        0K    0K sleep   99  -20   0:00:15 0.0% kmem_task/1
   501 root     9760K 1436K sleep   59    0   0:00:00 0.0% automountd/4
   499 root     9668K 1360K sleep   59    0   0:00:00 0.0% automountd/2
   488 root       14M 3896K sleep   59    0   0:00:00 0.0% inetd/4
   479 root     2780K 1488K sleep   59    0   0:00:00 0.0% hotplugd/3
   487 root     8928K 1164K sleep   59    0   0:00:00 0.0% cron/1
  1817 root       16M 3656K sleep   59    0   0:00:00 0.0% sshd/1
   468 daemon   7268K 4648K sleep   59    0   0:00:00 0.0% statd/1
   415 daemon   3508K 1440K sleep   59    0   0:00:02 0.0% rpcbind/1
share|improve this question
    
What sample interval ? Can you post vmstat and prstat outputs ? –  jlliagre May 19 '12 at 14:13
    
just the default output of top. I've put the outputs for vmstat and prstat into the post. However, right now the system seems to be ok. –  Pavel May 19 '12 at 14:24
1  
Yes vmstat/prstat statistics collected during the periods with the high value would help. What is the freeze duration ? –  jlliagre May 19 '12 at 15:48
3  
That's a different story. You aren't that much asking about hight values in kernel statistics but more about OS hangs. What you need is to force a crash dump when the freeze starts and analyze it later. docs.oracle.com/cd/E23824_01/html/821-2731/… –  jlliagre May 20 '12 at 20:52
1  
The first vmstat data line is at best an average since last boot. You should post the lines following that one for a better picture. –  jlliagre May 24 '12 at 7:00
show 2 more comments

1 Answer

This looks like a hardware problem with the disk, disk controller, SCSI/SAS cables or a software problem in ZFS.

You should open a case with Oracle.

If the server is completely frozen it is possible to generate a memory dump from OpenBoot.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.