Re: Update of table lags execution of statement by >1 minute? - Mailing list pgsql-general

From Tom Lane
Subject Re: Update of table lags execution of statement by >1 minute?
Date
Msg-id 12965.1184797481@sss.pgh.pa.us
Whole thread Raw
In response to Re: Update of table lags execution of statement by >1 minute?  (Erik Peterson <epeterson@cardiomems.com>)
List pgsql-general
Erik Peterson <epeterson@cardiomems.com> writes:
> I=B9m not sure the corrupted index issue is it.  After updating, the attribut=
> e
> shows up as the =B3old=B2 value with selects on different columns, ie:

> UPDATE mytable SET myattribute=3D1 WHERE id=3D14;
> COMMIT;
> SELECT * from mytable WHERE myattribute=3D0

> Would include the that row (id=3D14).  So it isn=B9t just a single corrupted
> index, if that is indeed the issue.

Hm.  I still think there's something going on in your application that
you're missing.  While you are "watching" the row to see it update,
would you select the system columns ctid,xmin,xmax as well as the data?
That might provide some clue what's really happening.

            regards, tom lane

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Feature request: Per database search_path
Next
From: "Pat Maddox"
Date:
Subject: Re: Need help optimizing this query