Re: Logical Decoding Failover - Mailing list pgsql-general

From Colin Morelli
Subject Re: Logical Decoding Failover
Date
Msg-id CAPtU-UrN5qE9_zai6e-TDGTMaVw9g=YwmL7Hz4aWvwDoPrG1Hw@mail.gmail.com
Whole thread Raw
In response to Re: Logical Decoding Failover  (Venkata Balaji N <nag1010@gmail.com>)
Responses Re: Logical Decoding Failover
List pgsql-general
Venkata,

Thanks for the reply. Unfortunately something like PgPool still won't create the replication slots on all hosts, and record the LSN in a way that is reusable on the secondary.

This sort of puts logical decoding at odds with HA, correct? In the case of master failover, there's no way to: a) know where in the stream you read to, or b) convert that to something that can be understood by a replica that's about to be promoted.

Are there any scheduled improvements here either for 9.6 or 9.7?

Best,
Colin
On Sun, Aug 7, 2016 at 1:50 AM Venkata Balaji N <nag1010@gmail.com> wrote:
On Sat, Aug 6, 2016 at 1:17 PM, Colin Morelli <colin.morelli@gmail.com> wrote:
Hey all,

I'm using logical decoding in my application to capture change streams and ship them to Kafka. However, logical replication slots aren't included in the WAL and thus don't make it to replicas. In the case of a failover, it's not clear what (if anything) can be done to continue receiving a change stream.

I noticed there was a failover slot patch that looks like it won't be making it for 9.6 (if at all), but does anyone else have suggestions for handling failover + logical decoding? Is there an answer, or is it limited to being a very manual process for now?

Yes, as of i know, It is a manual process for now. Another alternative you can consider (if possible at all) would be a tool like PgPool-II which is actually used for connection pooling, load balancing and automatic failover. It may not make sense if you are do not wish to pool you connections.

Regards,
Venkata B N

Fujitsu Australia

pgsql-general by date:

Previous
From: Михаил
Date:
Subject: ФС против Постгреса для файлов
Next
From: Aviel Buskila
Date:
Subject: fixing failed master after standby promotion