Re: strange behavior (corruption?) of large production - Mailing list pgsql-hackers

From Joe Conway
Subject Re: strange behavior (corruption?) of large production
Date
Msg-id 4390EC19.2090006@joeconway.com
Whole thread Raw
In response to Re: strange behavior (corruption?) of large production database  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:
> Joe Conway <mail@joeconway.com> writes:
> 
>>Since this database has many large, but static tables (i.e. new data is 
>>loaded each day, but the tables are partitioned into year-month tables), 
>>I'm thinking we can run VACUUM FREEZE on the whole database once, and 
>>then run VACUUM FREEZE periodically on just the tables that have had 
>>rows added since the last time -- is that correct?
> 
> That would work for the user tables, but the lesson to draw from this is
> not to forget about the system catalogs ...
> 

Right, I was just thinking about that too. Important safety tip :-)

Thanks again!

Joe



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: strange behavior (corruption?) of large production database
Next
From: "Charles F. I. Savage"
Date:
Subject: Building Windows Server Extensions Using VC++ 2005