Tim Kane wrote:
> Is anyone aware of cases where psql will occasionally return an exit status 
> of 1, despite there being
> no obvious error condition?
> 
> I have a regular (daily) cron that executes a fairly simple script of the 
> form:
> 
>       psql --set ON_ERROR_STOP=1 -h $DB_HOST -U $DB_USER -a -v 
> DATE="'${DATE}'" -v
> YYYYMM="${DATE_YYYYMM}" -f duplicate_removal.sql $DB_NAME >> $DB_LOG 2>&1
> 
> The execution of this duplicate_removal.sql script appears to run without 
> error. It’s fairly simple,
> and never fails to execute in and of itself.
> However, on very rare occasion, the psql session will run to completion, 
> without visible/logged error
> – but it returns an exit status 1.

> I can see that the output is correctly redirected to $DB_LOG and I can see 
> the successful COMMIT
> response returned at the end of the session. I can also verify that the 
> behaviour of the script is as
> expected, to completion.
> 
> The psql man page suggests that an exit status of 1 is a fatal error of its 
> own.
> 
>       EXIT STATUS
>              psql returns 0 to the shell if it finished normally, 1 if a 
> fatal error of its own occurs
>            (e.g. out of memory, file not found), 2 if the connection to the 
> server went bad and
>            the session was not interactive, and 3 if an error occurred in a 
> script and the variable
>            ON_ERROR_STOP was set.
> 
> There is no indication in the postgres log file of anything adverse.
> 
> At this point my only option is to set log_statement = all, enable shell 
> debugging, and wait a few
> months for it to happen again (I’ve seen this occur only twice, over a 2-3 
> month period).  Has anyone
> encountered this behaviour before?

According to the documentation you quoted, it would be useless to set
log_statement=all, because the problem is on the client end.

Shell debugging sounds slightly more promising, but it will probably
not tell you much more than that psql returned with exit code 1.

A quick look at the code gave me the impression that psql will always
write an error message before exiting with 1, but I may have missed a case.

Are you sure that there is nothing on stdout or stderr in the cases where
psql returns 1?

What is the last psql command that is executed?

Yours,
Laurenz Albe

-- 
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

Reply via email to