On Tue, Jan 25, 2011 at 7:20 PM, Filipe David Manana <[email protected]> wrote: > On Tue, Jan 25, 2011 at 6:16 PM, Benoit Chesneau <[email protected]> wrote: >> >> I would suggest 3 or 5 retry. because at the end it make a process >> that waited for >> >> 5 + 5000 + 50000 + 500000 .... > > Benôit, I think you didn't get it right. It's like: > > 5, 10, 20, 40, 80, ... > Yes asynchronicity will play here and my mail fixing that comming for you after you wrote this one :)
> It doubles. > I'll make the number of retries configurable, good suggestion. However > need a good name for the parameter, as people might think > "max_retries" is related to the number of retries per connection and > not the number of retries to trigger a replication by the replication > manager. > > replication_retries ?
