| From: | Craig Ringer <craig(at)2ndquadrant(dot)com> |
|---|---|
| To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
| Cc: | PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: WIP: Failover Slots |
| Date: | 2016-01-20 14:59:15 |
| Message-ID: | CAMsr+YEmoZhyGgtB+BK7s08BzzEfuBC_2Z7C5JtfFkf-NyLXLw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 20 January 2016 at 21:02, Craig Ringer <craig(at)2ndquadrant(dot)com> wrote:
> TL;DR: this doesn't work yet, working on it.
>
Nothing is logged on slot creation because ReplicationSlot->data.failover
is never true. Once that's fixed by - for now - making all slots failover
slots, there's a crash in XLogInsert because of the use of reserved bits in
the XLogInsert info argument. Fix pushed.
I also noticed that slot drops seem are being logged whether or not the
slot is a failover slot. Pushed a fix for that.
The WAL writing is now working. I've made improvements to the rmgr xlogdump
support to make it clearer what's written.
Slots are still not visible on the replica so there's work to do tracing
redo, promotion, slot handling after starting from a basebackup, etc. The
patch is still very much WIP.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Alvaro Herrera | 2016-01-20 15:16:24 | Re: checkpointer continuous flushing |
| Previous Message | Konstantin Knizhnik | 2016-01-20 14:55:09 | Re: Batch update of indexes |