[Bucardo-general] Should Bucardo Treat Deadlocks as Serialization Failures?

David E. Wheeler david at justatheory.com
Sat Nov 3 22:03:23 UTC 2012


On Nov 3, 2012, at 12:06 PM, Greg Sabino Mullane <greg at endpoint.com> wrote:

> Makes sense to me. I implemented it just now in 0406ecb5993793b59fa84815c2b0384979c862b4 
> but kept it as a separate knob, kid_deadlock_sleep, just in case.

Looks good, though a bit redundant. I think there is now a bunch of other deadlock stuff that can be ripped out, too. Oh, and I think `upgrade` needs to add a row for the new setting.

I can take care of this stuff on Monday (as well as test changes, if need be) unless you beat me to it. :-)

This is great!

Thanks,

David



More information about the Bucardo-general mailing list