Administration

[Dineshkumar]

Hi All,

Please advice me in MCSG environment

what can be done as Day to day administration on MC Service guard

Please advice me ?

Regards
Dinesh

[A. Clay Stephenson]

Ideally almost nothing. You should have monitoring systems in place (e.g. OpenView Operations) that constantly monitor the systems for hardware failures, network failures, application/package failures, resource limit approaches, security issues, monitor backups, ... . These messages are then forwarded to the appropriate party for action --- all automatically. You are really only limited by your own skill as to what can be done. For example, on some of my filesystems, if OV/O detects that a filesystem is 90% full, it will automatically extend the LVOL and grow the filesystem "on the fly" and my only interaction is a notification that the operation was done.

You should also be throughly comfortable with rotine tasks such as disk replacement "on the fly".

Ideally, in addition, to a test environment, you will also have a Sandbox cluster that you can use to "play"; ie, workout your hardware replacement procedures, software upgrade procedures, etc. without disrupting either the Test or Production environments.

When MC/SG is done right, it does almost nothing because all the problems (network failures, disk failures, etc.) are automatically handled long before MC/SG ever comes into play. You see, when you get your systems and procedures robust enough for MC/SG to work well, MC/SG does almost nothing --- and that is why you buy MC/SG. I have literally never had a package failover that I did not deliberately induce in many years of MC/SG operation.

[Dineshkumar]

Clay have answered my queries !!