hi,

why we need to bounce wm IS? If so, what are the steps we have to follow before bouncing IS?

regards,
Srinivas.

Hi,

There can be many reasons for restarting the IS.Few of them can be IS is unresponsive,if services are hung, some patches if applied need a restart of the server.

HTH

Who says we need to bouce IS !!

you are lucky that your systems are so stable that you don’t see/know the reason to “why it should be bounced”. Keep it up.

Better you don’t know the reason if you really don’t have to bounce your IS. If at all you really need to bounce your IS, better know the reason of bouncing it. Bouncing without knowing the reason will not ensure that you have will to have bounce it again or not. If at all you have to bounce it again make sure you take a back up of your unbounced IS. :lol::lol::lol::lol:

Hope I am not confusing you.

Hi jitengupta,

You got confused with my question …

what i was asking is that, if at all I want to bounse IS server, I want to know what are the steps/precautions i have to take before doing that. Just asking for an example when we will require IS to be bounced.

regards,
Srinivas.

In general, you’ll want to consider these steps to take an IS instance out of service in a controlled manner so that you don’t abort any in-progress services:

• Suspend all notifications (JDBC adapter, and others)
• Suspend Broker document retrieval
• Disable all ports, except an administrative-only port so that you can re-enable things (and do the planned maintenance) – this also generally means that the “normal traffic” port, 5555, should not be the primary port since it cannot be disabled.
• Suspend all scheduled tasks

These steps will “disconnect” IS from the world, preventing new activity from starting. Then you can wait a short period to let the running services complete and then do whatever maintenance necessary.

While it may not be necessary to do all these steps (for example, if you plan work during off-hours and nothing is going on, then they can probably be safely skipped) they can be useful in preventing failed integrations caused by interrupting something that was in progress.

These are all doable using IS Administrator. In the past I’ve created helper services that can do these steps via one service invocation, but it involves using undocumented services and APIs. It would be nice if wM put something together that would do this sort of orderly disconnect and shutdown after all in-progress work is complete.

Some times Server will not respond to the actions we do like disable notifications etc… if the server is busy or having problems, then the best option would be check the service usage if nothing is running use kill command on unix.

That’s true so an orderly shutdown isn’t possible. The list above is normally for planned outages (maintenance windows, etc.) or for unplanned outages when the server is still responsive.

Hello Rob,
Can you please let us know about the helper services that can be used before planned server bounce.

Thanks in Advance for your help.

Regards,
Soumodip

I mentioned in my post that all the steps can be accomplished using IS Administrator. You can look at the services that those pages use to find out which services you might invoke from your own helper service.