Thread: Relation does not exist

Relation does not exist

From
"Rob Klaus"
Date:
Hello,

We are having a problem where intermittently a query will not work,
returning an error:

ERROR:  Relation "x" does not exist

It will do this for a variable period, typically not lasting more than 5
minutes

I have searched through list archives and only find this reference in
regards to entities that are capitalized, which is *not* the case here.

Our setup may be 'bad', but I'm not sure; here is some background:

I have one database with 13 schemas
The code used to access data for the most part does not use schema.entity
naming convention
as we have the search_path set in postgresql.conf set to include all 13
schemas.

Entity names are not duplicated across schemas.

If I use the qualified schema.entity in my database calls, I don't get the
error.

I can go through all the code used to access data and add the schema name as
appropriate, but would prefer not to.

Has anyone seen an issue like this.  We have had this setup for about a year
without problems until just last week.

We are using PG_VERSION 7.3 on RedHat 7.0
System has 512 Mb RAM

I have tried tweaking some of the memory settings thinking maybe there was
an issue with that. . . .here are the details of that:

max_connections = 150
shared_buffers = 15200
sort_mem = 32168
fsync = false

We did change RedHat's shared memory setting to: kernel.shmmax = 134217728

Not sure if any other information would be useful.  I appreciate any
responses.

Thanks!

-Rob

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rob Klaus
Manager of Information Systems
Unique Management Services, Inc
119 East Maple Street
Jeffersonville, IN 40206
800.879.5453
http://www.unique-mgmt.com






Re: Relation does not exist

From
"Rob Klaus"
Date:
In general, it seems to correspond to when there are a lot of inserts being
run.  I have one application that maybe does 10 inserts within 15 seconds or
so, at which point the problem seems to crop up.

Any operation that does not use the qualified schema.entity names will error
out (selects, inserts, updates)

Taking your advice on upgrading both OS and pg version.  I'll be back if
that does not solve the issue ;-)

Thanks for your time.

-Rob

-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
Sent: Tuesday, August 10, 2004 2:30 PM
To: Rob Klaus
Cc: pgsql-general@postgresql.org
Subject: Re: [GENERAL] Relation does not exist


"Rob Klaus" <rob@unique-mgmt.com> writes:
> We are having a problem where intermittently a query will not work,
> returning an error:
> ERROR:  Relation "x" does not exist
> It will do this for a variable period, typically not lasting more than 5
> minutes

This certainly seems like a bug, but I have no idea what's causing it.
Can you characterize the conditions under which the bug appears or
disappears?  What operation exactly is getting the error, and what sorts
of things are happening in parallel with it?

> We are using PG_VERSION 7.3 on RedHat 7.0

Hm.  Red Hat 7.0 is positively ancient; you really ought to update.  And
if you are running original PG 7.3 there are some very serious bugs that
you ought to fix by updating to 7.3.6 (if not 7.4.something).  But I
can't say that I recognize your report as related to any of the known
bug fixes.

            regards, tom lane



Re: Relation does not exist

From
Tom Lane
Date:
"Rob Klaus" <rob@unique-mgmt.com> writes:
> We are having a problem where intermittently a query will not work,
> returning an error:
> ERROR:  Relation "x" does not exist
> It will do this for a variable period, typically not lasting more than 5
> minutes

This certainly seems like a bug, but I have no idea what's causing it.
Can you characterize the conditions under which the bug appears or
disappears?  What operation exactly is getting the error, and what sorts
of things are happening in parallel with it?

> We are using PG_VERSION 7.3 on RedHat 7.0

Hm.  Red Hat 7.0 is positively ancient; you really ought to update.  And
if you are running original PG 7.3 there are some very serious bugs that
you ought to fix by updating to 7.3.6 (if not 7.4.something).  But I
can't say that I recognize your report as related to any of the known
bug fixes.

            regards, tom lane