Re: BDR node removal and rejoin

On 14 July 2017 at 00:09, Zhu, Joshua <jzhu(at)vormetric(dot)com> wrote:

>
>
> Found these log entries from one of the other node:
>
>
>
> t=2017-07-13 08:35:34 PDT p=27292 a=DEBUG: 00000: found valid replication
> identifier 15
>
> t=2017-07-13 08:35:34 PDT p=27292 a=LOCATION:
> bdr_establish_connection_and_slot, bdr.c:604
>
> t=2017-07-13 08:35:34 PDT p=27292 a=ERROR: 53400: no free replication
> state could be found for 15, increase max_replication_slots
>
>
>
> Increased max_replication_slots, things are looking good now, thanks.
>
>
>
> This does bring up a couple of questions:
>
>
>
> 1. Given the fact there is no real increase in the number of nodes in
> this repeated removal/rejoining exercise, yet it caused replication slots
> being used up, wouldn’t removal of a node also automatically free up the
> replication slot allocated for the node?
>
>
Yes, it should. Open issue. A patch would be welcomed.

>
> 1. Or is there a way to manually free up no longer needed slots? (the
> don’t seem to show up in pg_replication_slots view, I made sure to use
> pg_drop_replication_slot when they do show up there)
>
>
It'll be complaining about replication identifiers ("origins" in 9.6); see
pg_replication_identifier

> 1. If there is such a thing, what is the rule of thumb for best value
> of max_replication_slots (are they somehow related to the value
> max_wal_senders as well), with respect to, say, the max number of nodes
> intended to support?
>
>
I think that's covered in the docs, but it's safe to err fairly high. The
cost of extra slots is minimal.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services