[racktables-users] Re: Using racktables with Rocks clusters
- From: Adrian Sevcenco <Adrian.Sevcenco@xxxxxxxxxxxxxxx>
- To: racktables-users@xxxxxxxxxxxxx
- Date: Fri, 11 Dec 2015 10:09:47 +0200
On 10/27/2015 12:13 PM, Sill, Alan wrote:
We run a medium-sized set of university-based data centers in three
locations (total about 45 racks and 15K cores) and are thinking of
using racktables as part of our internal organization and management
scheme.
A fairly large fraction of what we do is based on the Rocks cluster
management suite. Is there any experience on this list in using
racktables with Rocks clusters? A plug-in or adapter to import
racktables descriptions from the Rocks database would seem to be a
fairly natural starting point, but we don’t want to duplicate this
work if it has already been done.
I would also be interested in this :) anyway the best bet would be
not the rocks database (witch can change as it changed from 6.1.1 to 6.2)
but the snmp services that are running on all nodes ..
So, assuming that with a little bit of scripting you gather from all snmpd
information on the cluster nodes i would have the following questions :
1. is there a possibility to feed a file with information to racktables that
will be parsed and the objects added? if yes what format is needed?
2. i seen some configurations about cacti and munin .. what they do? can
something
similar be added for ganglia? maybe to each server to be able to add a link
to the ganglia host page of that server?
Thank you!!!
Adrian
Also any hints on discovery or automatic detection (e.g., using arp)
of existing configurations would be useful. Tips for newcomers would
be appreciated!
Alan
--
----------------------------------------------
Adrian Sevcenco, Ph.D. |
Institute of Space Science - ISS, Romania |
adrian.sevcenco at {cern.ch,spacescience.ro} |
----------------------------------------------
Other related posts:
