From: | Andres Freund <andres(at)anarazel(dot)de> |
---|---|
To: | Craig Ringer <craig(at)2ndquadrant(dot)com> |
Cc: | Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Assertion failure when streaming logical changes |
Date: | 2015-04-07 12:54:20 |
Message-ID: | 20150407125420.GA12291@awork2.anarazel.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 2015-04-07 17:22:12 +0800, Craig Ringer wrote:
> It might be a good idea to apply this if nothing better is forthcoming.
> Logical decoding in WALsenders is broken at the moment.
Yes.
> Otherwise it needs to go on the 9.5 blockers list.
>
> But could we get rid of those active_count manipulations too? Could you
> > replace the SnapBuildSnap[Inc|Dec]Refcount calls with
> > [Un]RegisterSnapshot()?
> >
>
> It would be interesting to know why it was done that way in the first
> place, rather than using the snapshot management infrastructure.
Because the snapshot tracking infrastructure isn't particularly
suitable. These snapshots are much longer lived; they span transactions
and such. Nothing snapmgr.c is made for. It seemed more
complicated/bug-prone to change snapmgr.c to support a foreign use case.
Greetings,
Andres Freund
From | Date | Subject | |
---|---|---|---|
Next Message | Andres Freund | 2015-04-07 13:08:03 | Re: PATCH: Add 'pid' column to pg_replication_slots |
Previous Message | Craig Ringer | 2015-04-07 12:11:49 | Row security violation error is misleading |