error messages in syslog and in dmesg

[VINU]

hello friends,
I am getting the following error in syslog.log,

May 10 20:58:31 suboz024 vmunix: sysmap: rmap ovflo, lost [91061l,91062l)

and when I run 'dmesg' it is giving only the following output,

suboz024:/ > dmesg
May 10 23:45
Magic number wrong (namelist mismatch?)
suboz024:/ >

Could anyone tell me what this error means

thanks & regards
vinu

[A. Clay Stephenson]

rmap ovflo means that the resource map has overflowed. The resource map is a kernel data structure used to keep track of kernel virtual memory allocation. Since you didn't bother to indicate the OS version, it's difficult to suggest patches. Search the patch database and use the string "rmap" to search on. It is also possible that you have a severe memory shortage.

[hari jayaram_1]

Vinu,

The rmap ovflo is due to a application memory leak. There was a pretty good discussion about this on the itrc forum sometime back and the link is below
http://forums1.itrc.hp.com/service/forums/questionanswer.do?admit=716493758+1084225919365+28353475&threadId=70397

For the second message again here is the lnk below

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xbabf0bce6f33d6118fff0090279cd0f9,00.html

[Michael Tully]

You may have a memory leak in your application. Check this posting and link.

http://forums1.itrc.hp.com/service/forums/parseCurl.do?CURL=%2Fcm%2FQuestionAnswer%2F1%2C%2C0xebf3911284f5d5118ff40090279cd0f9%2C00.html&admit=716493758+1084226557584+28353475

[Jairo Campana]

SOLUTION SEE:
http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062922113

ALSO:

http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000072951466

[VINU]

dear friends,
thanks for ur valuable inputs,it helped me a lot.
one more question.can we check the programs with memory leak using commands/CSTM.


regards
vinu

[A. Clay Stephenson]

You can run ps commands every 30 seconds or so and log the output.

typeset -i10 KNT=1
while [[ ${KNT} -le 100 ]]
do
echo "${KNT}" >> mylog
date >> mylog
UNIX95= ps -e -o vsz,comm >> mylog
echo "" >> mylog
sleep 30
(( KNT += 1 ))
done

[Don Morris_1]

The resource map in question being the sysmap -- this is a fragmentation of the Kernel virtual address space. This is usually caused by a packet storm or other behavior where the kernel does a lot of small allocations which are then non-contiguously freed.

If you're running 11.0 or later (I don't think this was on 10.20), try doubling the "nsysmap" tunable. This is exactly the problem it was introduced to address.

Regarding the dmesg output -- I usually see this when I boot from a kernel other than /stand/vmunix -- dmesg is telling you that /dev/kmem and /stand/vmunix don't match up. If you booted from a different kernel (say /stand/vmunix_tmp), run it as "dmesg /dev/kmem /stand/vmunix_tmp".