Is there any way to monitor disk I/O stats on UNIX servers?
There are potentially two ways to do this:
If the Unix systems have an snmp agent that reports this information you can collect, sore and report on it using the Universal Device Poller.
Your other option would be to use the Orion APM and run a script against the system that collects that information as a statistic for you.
I am also looking to do the same thing. We have Solarwinds communicating via SNMP ok to the Unix server. I can see the unix server and it reports network interface util. But no CPU, MEM, disk volume info/stats. Any ideas. I've been told from a unix tech that there should be no reason that we should not be able to report this data via SNMP?
Richard.
Rich,
You need to determine which MIBs are supported by the platform and specific UNIX OS your systems are running. Once you know that you can use UnDP to create polls to monitor any metric within those MIBs. I would assume your "unix tech" could provide this info but if not and given that you have SNMP comm. to the server you could do an SNMP Walk of the server to see which MIBS respond.
We use APM Linux/Unix Script Monitor -
For Reads Mb/s
iostat <Device> -d -x -m | grep <Device>| sed 's/[ \t]\+/ /g' |cut -f6 -d " "
For Writes Mb/s
iostat <Device> -d -x -m | grep <Device> | sed 's/[ \t]\+/ /g' |cut -f7 -d " "
Replace <Device> with what you want to monitor
JB
