cpu metrics on SolarisHi all,
We have installed ganglia at SUN 6800(SunOS 5.9 + ia64), when we run gmond,
it display:
gmond: cannot open /dev/kmem
kvm_open: Permission denied
*** kvm_open() failed, are you running gmond as root?
kvm_nlist: Bad address
Segmentation Fault (core dumped)
What's wrong with it? Thanks for any tips/suggestions/help.
Sincerely Yours
Wenguo WEI
[EMAIL PROTECTED]
[EMAIL PROTECTED]
-----原始邮件-----
发件人: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] [EMAIL PROTECTED]
发送时间: 2004年4月7日 16:14
收件人: [email protected]
主题: [Ganglia-general] cpu metrics on Solaris
Hi all,
Just wanting to include a Solaris 8 machine to my ganglia monitoring... So I
retrieved and compiled the sources. No big deal up to there but when I start
the gmond daemon, first it fail and exit (no error code) when I do not start it
in debug mode (level 1 is enough). But then it seems to have problem collecting
the cpu values! The other values seem to be collected ok.Did someone already
see and solve this problem?
To me it look like that the picking of values in solaris.c is off the mark
but my programming skills are far away :( and maybe the problem is elsewhere!
When I start gmond with debug level 1, I got this answer:
cpustuff: Error 0
cpustuff: Error 0
cpustuff: Error 0
cpustuff: Error 0
cpustuff: Not enough space
cpustuff: Not enough space
cpustuff: Not enough space
cpustuff: Not enough space
At debug level 2, I got this (i cut a lot):
offset = 144515136, cpu_now[1] = 535191074
cpustuff: Not enough space
offset = 144509640, cpu_now[1] = 604166314
cpustuff: Not enough space
offset = 272742264, cpu_now[1] = 697194745
Raw: bread / bwrite / lread / lwrite / phread / phwrite
16369260,16369253,7 / 99529321,99529179,142 / 3841781482,3841772867,8615 /
656923319,656919614,3705 / 502596970,502596966,4 / 1426660978,1426660978,0
Aftermath: 0.388889 7.888889 478.611115 205.833328 0.222222 0.000000 delta =
18
** ** ** ** ** Are percentages electric? Try 1000%, 0% , 0% , 0% , 0% 0%
set_metric_value() exec'd mem_buffers_func (22)
Regards,
Laurent