Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-19 Thread Alvaro Herrera
Tom Lane wrote: Josh Berkus j...@agliodbs.com writes: On 03/17/2014 05:49 PM, Josh Berkus wrote: https://wiki.postgresql.org/wiki/20140320UpdateIssues Anyone? We're going public with this in 36 hours, so I'd love for it to actually be correct. I did a bit more hacking on this page.

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-19 Thread Josh Berkus
On 03/19/2014 10:37 AM, Alvaro Herrera wrote: Tom Lane wrote: Josh Berkus j...@agliodbs.com writes: On 03/17/2014 05:49 PM, Josh Berkus wrote: https://wiki.postgresql.org/wiki/20140320UpdateIssues Anyone? We're going public with this in 36 hours, so I'd love for it to actually be correct.

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-19 Thread Tom Lane
Josh Berkus j...@agliodbs.com writes: On 03/19/2014 10:37 AM, Alvaro Herrera wrote: I wonder about suggesting other versions of ALTER TABLE that can do heap rewrites. I don't want to recommend any version of ALTER TABLE until someone actually tests it on a corrupted database. A test would

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Josh Berkus
On 03/17/2014 05:49 PM, Josh Berkus wrote: All, https://wiki.postgresql.org/wiki/20140320UpdateIssues I'm sure my explanation of the data corruption issue is not correct, so please fix it. Thanks! Anyone? We're going public with this in 36 hours, so I'd love for it to actually be

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Greg Stark
On Tue, Mar 18, 2014 at 6:41 PM, Josh Berkus j...@agliodbs.com wrote: Anyone? We're going public with this in 36 hours, so I'd love for it to actually be correct. I'll do a first pass now. -- greg -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Greg Stark
On Tue, Mar 18, 2014 at 7:05 PM, Greg Stark st...@mit.edu wrote: I'll do a first pass now. I fixed the Causes section. I haven't touched the other sections which are pretty reasonable. It would be nice to have more suggestions for what people should do other than dump/restore. It would be nice

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Andres Freund
On 2014-03-18 19:28:53 +, Greg Stark wrote: It would be nice to be able to tell people that if they vacuum, then reindex and check all their foreign key constraints then they should be ok. I don't think so:

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Josh Berkus
On 03/18/2014 12:35 PM, Andres Freund wrote: I still think a rewriting noop ALTER TABLE ... ALTER COLUMN col TYPE old_type USING (col); is the only real thing to do. Then why wouldn't VACUUM FULL work? -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com -- Sent via pgsql-hackers

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Andres Freund
On 2014-03-18 12:52:49 -0700, Josh Berkus wrote: On 03/18/2014 12:35 PM, Andres Freund wrote: I still think a rewriting noop ALTER TABLE ... ALTER COLUMN col TYPE old_type USING (col); is the only real thing to do. Then why wouldn't VACUUM FULL work? Please read the referenced message and

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Josh Berkus
On 03/18/2014 12:55 PM, Andres Freund wrote: On 2014-03-18 12:52:49 -0700, Josh Berkus wrote: On 03/18/2014 12:35 PM, Andres Freund wrote: I still think a rewriting noop ALTER TABLE ... ALTER COLUMN col TYPE old_type USING (col); is the only real thing to do. Then why wouldn't VACUUM FULL

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Tom Lane
Andres Freund and...@2ndquadrant.com writes: On 2014-03-18 19:28:53 +, Greg Stark wrote: It would be nice to be able to tell people that if they vacuum, then reindex and check all their foreign key constraints then they should be ok. I don't think so:

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Andres Freund
On 2014-03-18 16:19:01 -0400, Tom Lane wrote: Andres Freund and...@2ndquadrant.com writes: On 2014-03-18 19:28:53 +, Greg Stark wrote: It would be nice to be able to tell people that if they vacuum, then reindex and check all their foreign key constraints then they should be ok. I

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Alvaro Herrera
Andres Freund wrote: On 2014-03-18 16:19:01 -0400, Tom Lane wrote: Andres Freund and...@2ndquadrant.com writes: On 2014-03-18 19:28:53 +, Greg Stark wrote: It would be nice to be able to tell people that if they vacuum, then reindex and check all their foreign key constraints then

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Andres Freund
On 2014-03-18 17:34:34 -0300, Alvaro Herrera wrote: Andres Freund wrote: On 2014-03-18 16:19:01 -0400, Tom Lane wrote: Andres Freund and...@2ndquadrant.com writes: On 2014-03-18 19:28:53 +, Greg Stark wrote: It would be nice to be able to tell people that if they vacuum, then

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Andrew Dunstan
On 03/18/2014 04:39 PM, Andres Freund wrote: Mail that's CC/TOed to me onlist, is automatically marked as read by a sieve script so I don't have to mark it as read twice. It seems something went wrong there for a couple of messages... Why not just turn on eliminatecc on the majordomo

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-18 Thread Tom Lane
Josh Berkus j...@agliodbs.com writes: On 03/17/2014 05:49 PM, Josh Berkus wrote: https://wiki.postgresql.org/wiki/20140320UpdateIssues Anyone? We're going public with this in 36 hours, so I'd love for it to actually be correct. I did a bit more hacking on this page. Could use another look

[HACKERS] Wiki Page Draft for upcoming release

2014-03-17 Thread Josh Berkus
All, https://wiki.postgresql.org/wiki/20140320UpdateIssues I'm sure my explanation of the data corruption issue is not correct, so please fix it. Thanks! -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-17 Thread David Johnston
I sent a post to -general with a much more detailed brain dump of my current understanding on this topic. The main point I'm addressing here is how to recover from this problem. Since a symptom of the problem is that pg_dump/restore can fail saying that (in some instances) the only viable

Re: [HACKERS] Wiki Page Draft for upcoming release

2014-03-17 Thread David Johnston
Josh Berkus wrote All, https://wiki.postgresql.org/wiki/20140320UpdateIssues I'm sure my explanation of the data corruption issue is not correct, so please fix it. Thanks! I presume that because there is no way the master could have sent bad table data to the replication slaves that