Resent-From: POSTGRES mailing list <postman@postgres.Berkeley.EDU>
Resent-Message-Id: <199509040053.RAA09089@nobozo.CS.Berkeley.EDU>
Sender: owner-postman@postgres.Berkeley.EDU
Message-Id: <199509040054.AA19709@janus.cat.csiro.au>
To: sid@lamont.ldgo.columbia.edu (sidney hellman)
Cc: postgres@postgres.Berkeley.EDU
Subject: Re: backend problems and "copy to" 
In-Reply-To: Your message of "Tue, 29 Aug 1995 12:20:23 EDT."
             <9508291620.AA02134@lamont.ldgo.columbia.edu> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Mon, 04 Sep 1995 10:53:19 +1000
From: Nicholas Hudson <N.Hudson@janus.cat.csiro.au>
Resent-To: postgres-redist@postgres.Berkeley.EDU
Resent-Date: Sun, 03 Sep 95 17:53:40 -0700
Resent-XMts: smtp

Hi Sid,

> Hi there,
> 
> First of all, I'm not new to programming postgres, I'm just new to the
> nitty-gritty of administration. 
> 
> We area having some problems with our psotgres 4.2 database. 
> trying to back up one of the classes with the command "copy" from within
> monitor results in the following message. 
> 
> copy trace to "/passcal/data00/field/moma/moma.95:241:10:19.trace.ascii.new"
> \g
> 
> Query sent to backend is "copy trace to "/passcal/data00/field/moma/moma.95:241:10:19.trace.ascii.new" "
> Error: No response from the backend, exiting...

I'd like to make two statements on this. 

1. I have had problems restoring tables from a text copy. It seems to be a 
problem with abstime & reltime fields. Iwould suggest that you test this out 
in your development database before you trust it. I use the binary option and 
this works fine.

2. We get this problem from time to time on a couple of occations it was due 
to a coruption of some for in the database which I could not find or fix. (ie. 
I have very little exp. or knowledge about postgres db stucture). Most time 
the computer must be rebooted and the problem goes away, restarting the 
postmaster does nothing. Once I caused this symptom myself while doing some 
maintenance to a database I was making a small change to approx 1000 tuples in 
a class using a single transaction to protect the database integrity. 
Unfortunatly the tcl script crashed without doing an end or abort. Hay presto 
"Error: No response from the backend, exiting..." on everything. A reboot 
fixed this one, luckily. I get the impression that some table in the OS or 
something fills up ???

I havn't spent much time digging into the problem so sorry I don't have a 
solution for you. If any one can explain these symptoms I would be grateful.

> 
> This is after about 15 minutes, and the output file has grown to a size of
> about 16.5 MB (I did expect something this large, I just didn't expect
> an error.)
> 
> The same thing happens with something along the lines of this..
> 
> monitor -c "copy trace to stdout"  -NQ moma > filename.stuff
> 
> I decided to try to vacuum the DB, but the vacuum is failing with
> backend problems also.
> 
> 
> The DB still seems to be ok, from a users point of view.
> 
> I need to back up this db, and copy it to a machine at a remote
> sight. Does anyone have any ideas or insight?  Our normal DB guy is
> unreachable for the next few weeks, so I've kind of been tossed into
> the lion cage.

best of luck sid

Nicholas

 CSIRO dEM QCAT
 Computer Support
 N.Hudson@dem.csiro.au 
 Phone: +61 7 212 4604


==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/