Data Backup and Recovery

Watchdog timeout question.

vloiseau
3,270 Views

Hi,

I want to verify that I have understood the watchdog timeout and what occurs when the timeout is reached.

1st scenario :

Given a configuration with Domino, backup sequence is quiesce nsf, snapshot, unquiesce nsf

you are in the middle of the quiesce and the timeout is reached, does the watchdog ask for the unquiesce or kill the Process ?

But if the agent in monothreaded, how can he deal with the unquiesce request from the watchdog as he's dealing the previous quiesce request in one at a time manner ?

2nd scenario :

same configuration

Imagine you are in the middle of the unquiesce, what can do the watchdog if the timeout is reached ?

Thanks.

1 ACCEPTED SOLUTION

Arora_Kapil
3,270 Views

Scenario 1:

If quiesce is running and timeout is reached, nothing happens.

If quiesce is finished and no unquiesce is called and SC_AGENT_UNQUIESCE_TIMEOUT timeout is reached, unquiesce is called.

Scenario 2:

Watchdog doesn't do anything.

If you set SC_AGENT_WATCHDOG_ENABLE= Y then SC_AGENT_UNQUIESCE_TIMEOUT will determine the time that agent should wait before running the unquiesce on the config.

This is only to make sure agent doesn’t allow two queisce operations one after the other and if there is something wrong in communication between agent/server the database is brought back to normal state.

Thanks,

Kapil

View solution in original post

1 REPLY 1

Arora_Kapil
3,271 Views

Scenario 1:

If quiesce is running and timeout is reached, nothing happens.

If quiesce is finished and no unquiesce is called and SC_AGENT_UNQUIESCE_TIMEOUT timeout is reached, unquiesce is called.

Scenario 2:

Watchdog doesn't do anything.

If you set SC_AGENT_WATCHDOG_ENABLE= Y then SC_AGENT_UNQUIESCE_TIMEOUT will determine the time that agent should wait before running the unquiesce on the config.

This is only to make sure agent doesn’t allow two queisce operations one after the other and if there is something wrong in communication between agent/server the database is brought back to normal state.

Thanks,

Kapil

Public