Difference: ComputingAdminMLMonitoring ( vs. 1)

Revision 12015-04-27 - samir

Line: 1 to 1
Added:
> >	As of 2015, the only component maintained locally in the T2 is the temperature monitoring. It used to monitor a number of things in nodes that other systems such as Ganglia would do too. However the temperature monitoring is useful and still maintained. The whole system is a bit more complex than what it needed to be, it consists on the following components : PDUs with public IPs (???), a specific SNMPv2 community "newman" (the server) as a SNMP proxy. Not sure why. But one interesting function is that in /etc/snmp/snmp.conf one can see a map of OIDs, that also helps to find where is which PDU. t2-monitor - This is the final part of the loop. The actual ML component runs there and reads the information which is proxied by "newman". In order for the system to be up, all parts should work. One good way to test the first 2 (SNMP proxy, PDUs) is to snmpwalk them. Find examples in the history. If you can retrieve anything that makes sense you should be fine in that given step. The monitoring page for temperature/power is : http://monalisa4.ultralight.org/display?page=sensors/temp -- Main.samir - 2015-04-27

Line: 1 to 1

Added:

>
>

As of 2015, the only component maintained locally in the T2 is the temperature monitoring. It used to monitor a number of things in nodes that other systems such as Ganglia would do too.

However the temperature monitoring is useful and still maintained. The whole system is a bit more complex than what it needed to be, it consists on the following components :

PDUs with public IPs (???), a specific SNMPv2 community
"newman" (the server) as a SNMP proxy. Not sure why. But one interesting function is that in /etc/snmp/snmp.conf one can see a map of OIDs, that also helps to find where is which PDU.
t2-monitor - This is the final part of the loop. The actual ML component runs there and reads the information which is proxied by "newman".

In order for the system to be up, all parts should work. One good way to test the first 2 (SNMP proxy, PDUs) is to snmpwalk them. Find examples in the history. If you can retrieve anything that makes sense you should be fine in that given step.

The monitoring page for temperature/power is :

http://monalisa4.ultralight.org/display?page=sensors/temp

-- Main.samir - 2015-04-27

View topic | History: r1 | More topic actions...

Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback