Re: 10.2.0.3 RAC diag process - high cpu util
- From: "LS Cheng" <exriscer@xxxxxxxxx>
- To: jstegmair@xxxxxxxxxxx
- Date: Wed, 3 Dec 2008 18:28:09 +0100
That is probably a symptom.
Do you see more trace files? How is the database running? (Wait events,
Interconnect performance etc)
Thanks
--
LSC
On Wed, Dec 3, 2008 at 5:53 PM, Juergen Stegmair <jstegmair@xxxxxxxxxxx>wrote:
> Hi Everybody,
>
>
>
> Anyone had ever problems with a spinning/runaway diag process (rdbms
> instance, not asm) on 10.2.0.3 RAC / RH 4?
>
>
>
> > ls -altr *diag*
>
> ...
>
> -rw-rw---- 1 oracle oinstall 311506 May 10 2008
> prorac2_diag_27727.trc
>
> -rw------- 1 oracle oinstall 1096 May 14 2008
> prorac2_diag_13877.trc
>
> -rw------- 1 oracle oinstall 4985163676 Dec 3 10:27
> prorac2_diag_29143.trc
>
>
>
> > tail prorac2_diag_29143.trc
>
> SKGXPSEGRCV: MESSAGE TRUNCATED user data 48 bytes payload 120 bytes
>
> SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header
> 0x0x7fbfffd948 len 48 bytes
>
> SKGXPSEGRCV: MESSAGE TRUNCATED user data 48 bytes payload 120 bytes
>
> SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header
> 0x0x7fbfffd948 len 48 bytes
>
> SKGXPSEGRCV: MESSAGE TRUNCATED user data 48 bytes payload 120 bytes
>
> SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header
> 0x0x7fbfffd948 len 48 bytes
>
> SKGXPSEGRCV: MESSAGE TRUNCATED user data 48 bytes payload 120 bytes
>
> SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header
> 0x0x7fbfffd948 len 48 bytes
>
> SKGXPSEGRCV: MESSAGE TRUNCATED user data 48 bytes payload 120 bytes
>
> SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header
> 0x0x7fbfffd948 len 48 bytes
>
>
>
> And here's top output:
>
>
>
> top - 10:30:14 up 202 days, 11:16, 4 users, load average: 2.87, 2.77,
> 2.78
>
> Tasks: 369 total, 2 running, 367 sleeping, 0 stopped, 0 zombie
>
> Cpu(s): 24.3% us, 21.1% sy, 0.0% ni, 33.9% id, 19.2% wa, 0.0% hi, 1.4%
> si
>
> Mem: 16408900k total, 16379576k used, 29324k free, 176712k buffers
>
> Swap: 8388600k total, 677940k used, 7710660k free, 11023172k cached
>
>
>
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
>
> 29143 oracle 25 0 10.2g 91m 36m R 99.9 0.6 65151:03
> ora_diag_prorac2
>
> 31085 oracle 15 0 10.1g 1.3g 1.3g S 38.5 8.2 0:30.15 oracleprorac2
> (LOCAL=NO)
>
> 6137 oracle 16 0 10.2g 818m 798m S 13.3 5.1 0:10.59 oracleprorac2
> (LOCAL=NO)
>
> 2783 oracle 15 0 10.1g 2.6g 2.6g S 4.6 16.9 0:35.02 oracleprorac2
> (LOCAL=NO)
>
> 14250 oracle 15 0 10.1g 2.3g 2.3g S 3.0 14.9 0:27.06 oracleprorac2
> (LOCAL=NO)
>
>
>
> Killing the diag process just restarts it, and runs up to the high util
> again …
>
>
>
> ML search doesn't find anything, or any other searches.
>
>
>
> Any ideas?
>
>
>
> Thanks,
>
> Juergen Stegmair
>
> ENKITEC
>
Other related posts: