Subscribe

How to do a health check to see if the head is serving data to the user?

The theory is that any moment in the day by checking these stats he/she would have a reasonable accurate view about the filer status

Including but not limited to:

  • - Is the head serving data
  • - CPU / mem usage and problems
  • - Any possible problems / Network / CIFS etc...

I know DFM and possible Performance Advisor will help here so it might be a combination of some DFM and Command line checking.

Thank you.

Re: How to do a health check to see if the head is serving data to the user?

There are indeed several ways to do. I give just some examples that we use:

  1. Operation/performance Manager (aka dfm) is a great tool for this. OM can send alerts by a high cpu-use, host down, latencies, ...
  2. You can very easy write a script that checks some parameters and put those data in file, db, ...: we use it for reallocation measurement, the control of flexshare settings, ...
  3. We use also nagios: not only for just a ping check but it's also very easy to check if some services are running: cifs, nfs
  4. With syslog you can do a lot with the messages of the filer: we send all the new messages that we didn't marked as "not important for us" to the operations team.

reinoud

Re: How to do a health check to see if the head is serving data to the user?

In addition to the excellent suggestions made by Reinoud, for a long-term legacy viewpoint monitored internally could include.

Delivering ASUP's to an internal mail user for review/collection - Including notification if a critical event were to occur.

Also, outside of DFM/PA, if you happen to use any other SNMP monitor or traps, you could have it check the filer on specific triggers, though if something truly bad did happen, the ASUP alerts would then notify you.

Christopher