On 3/11/2014 5:50 AM, Aggarwal, Ajay wrote:
> Thats exactly what I was thinking after all other experiments. Couple
> of questions:
> 1) why did you say that 300 seconds is the upper limit? Is this
> enforced by Postgres? What if I want to set it to 10 minutes?
> 2) whats the downside of bigger replication timeout?
I said, set it to 300 or whatever. An unfortunate extra comma confused
my meaning. What I meant was, whatever you set it to, thats a upper limit.
As I understand it, that timeout is how long the replication can lag the
server before the server decides to stop replication. with it at 300,
under heavy load, the replication could run as much as 5 minutes (300
seconds) behind before it errors.
--
john r pierce 37N 122W
somewhere on the middle of the left coast