[pgpool-general: 1831] Re: AWS and pgpool-II

[David Kerr]

On Jun 5, 2013, at 5:52 PM, Tatsuo Ishii <ishii at postgresql.org> wrote:

>> On Wed, Jun 05, 2013 at 01:40:51PM +0900, Tatsuo Ishii wrote:
>> - Hi,
>> - 
>> - I have gotten a few complains about unwanted failover caused by TCP
>> - connection errors on AWS.
>> - 
>> - By googling "AWS pgpool", I found a similar case:
>> - 
>> - 1) https://forums.aws.amazon.com/thread.jspa?messageID=429906
>> - 
>> - Posted on Mar 3, 2014. Unfortunately in Japanese. In short, they got
>> - "No route to host" error periodically. The region is
>> - Tokyo(ap-northeast-1a). Up to now no replies so far.
>> - 
>> - I don't think Amazon is going to fix the problem soon. Maybe pgpool
>> - should be prepared to such an unstable envrionment. For now, I
>> - recommend to set fail_over_on_backend_error = off on AWS.
>> 
>> We're seeing that problem, it's immensly problematic for us. It doesn't
>> just affect PgPool / Postgres but can randomly hit any machine in your
>> VPC.
>> 
>> We've opened up a ticket with amazon support and have been working with
>> them on it. There is hope that we'll have a resolution soon.
> 
> Great. Could you please let us know if Amazon fixes the problem?

Definitely!