Horton

Re: 2pc leaks fds

Alvaro Herrera <alvherre@2ndquadrant.com>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Antonin Houska <ah@cybertec.at>
Cc: Andres Freund <andres@anarazel.de>, Michael Paquier <michael@paquier.xyz>, pgsql-hackers@postgresql.org
Date: 2020-04-21T23:54:22Z
Lists: pgsql-hackers 
On 2020-Apr-08, Antonin Houska wrote:

> Specifically for 2PC, isn't it better to close some OS-level FD of an
> unrelated table scan and then succeed than to ERROR immediately? Anyway,
> 0dc8ead46 hasn't changed this.

I think for full generality of the interface, we pass a "close" callback
in addition to the "open" callback.  But if we were to pass it only for
WALRead, then there would be no way to call it during XLogReaderFree.

I think the fix Andres applied is okay as far as it goes, but for the
long term we may want to change the interface even further -- maybe by
having these functions be part of the XLogReader state struct.  I have
this code paged out of my head ATM, but maybe tomorrow I can give it a
little think.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Commits

  1. Fix obsolete references to "XLogRead"

  2. Rework XLogReader callback system

  3. Fix XLogReader FD leak that makes backends unusable after 2PC usage.

  4. Refactor WAL file-reading code into WALRead()

  5. Add ArchiveOpts to pass options to ArchiveEntry