Hi Jason,
Assuming that Itanium gmond uses gmond/machines/linux.c, the number of CPUs
is retrieved using the glibc call "get_nprocs()" with the remainder of the
CPU info parsed from /proc/cpuinfo. You may want to check the source.
If you grep the gmond output (telnet one-of-your-monitoring-core-hosts
8649) for "cpu_num" you should be able to find out which one's reporting
the number of CPUs properly. :)
I don't know if we've got anybody doing ganglia development in an MP
Itanium environment. I don't recall any mention of it on the developer
list, anyway...
Since it's a glibc call, that implies this could be happening elsewhere...
I'm guessing it isn't but just in case, you may want to check /proc/cpuinfo
and your dmesg output to see if both processors are being reported by the
kernel.
[or 'all four,' if you're using a hyperthreading-enabled processor... can't
remember offhand whether the I2 is or not ... ]
Good luck...
Jason Holland wrote:
Hi,
I'm currently running ganglia 2.5.1 on an Itanium2 cluster to monitor
the nodes. The nodes are hp zx6000 workstations with dual 900 mhz
itanium2 cpu's. Ganglia is reporting an incorrect number of cpu's. For
the 24 workstations we have online right now, its reporting 25 cpu's. Has
anyone seen this before? How does ganglia determine this number? From
/proc? btw, we are running RedHat Advanced Workstation 2.1. Thanks
Jason P Holland
Texas Learning and Computation Center
http://www.tlc2.uh.edu
University of Houston
Philip G Hoffman Hall rm 207A
tel: 713-743-4850
cell: 281-451-5991
-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open!
Get cracking and register here for some mind boggling fun and
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
