[Drbd-dev] [CASE-24] Does replication have to wait until the end of handler service?

[Lars Ellenberg]

On Wed, Feb 24, 2016 at 04:04:11PM +0900, 김재헌 wrote:
> Hi,
> 
> Conf:
>  - A(Async) or C(sync) replication mode
>  - setup before-resync-source( this handler sleeps just 1 minute)
> 
> Test:
>  - 1) Big file copy
>  - 2) During copy,disconnect and connect
>  - 3) before-resync-source handler called
>  - 4) At this point copy operation is pending until handler finished.
> 
> Is it normal operation?

The before-resync handlers can return an exit code != 0 (and != 3,
for hysterical reasons), and will cause the connection to be dropped
(and nothing to be resynced or changed).

So yes, the resync has to wait for that exit code.

What would you want to do from there?

Also, if you think that is useful,
you can do the equivalent of "nohup some-long-running-script & exit 0"...
But I fail to see the usecase.


-- 
: Lars Ellenberg
: http://www.LINBIT.com | Your Way to High Availability
: DRBD, Linux-HA  and  Pacemaker support and consulting

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.