Thread: pg_dump failed with error code 255, but I don't see why

v9.6.2.23

I always run pg_dump with the --verbose option, and the bash script echos 
$?, so there's a 25,664 line log file for me to examine which explicitly 
shows the return code.

But however much I search the log file for strings like ERROR, FATAL, WARN, 
INFO, and "archiver", I don't see any problems.

What else should I look for?

-- 
Angular momentum makes the world go 'round.

Ron <ronljohnsonjr@gmail.com> writes:
> v9.6.2.23
> I always run pg_dump with the --verbose option, and the bash script echos 
> $?, so there's a 25,664 line log file for me to examine which explicitly 
> shows the return code.

A quick search of the 9.6 pg_dump source code shows that it only
ever calls exit() with values 0 or 1.  255 would have to be coming
from some outside factor.  You sure your shell script isn't
misbehaving and corrupting the reported return code?

            regards, tom lane

On 9/22/22 09:34, Tom Lane wrote:
> Ron <ronljohnsonjr@gmail.com> writes:
>> v9.6.2.23
>> I always run pg_dump with the --verbose option, and the bash script echos
>> $?, so there's a 25,664 line log file for me to examine which explicitly
>> shows the return code.
> A quick search of the 9.6 pg_dump source code shows that it only
> ever calls exit() with values 0 or 1.  255 would have to be coming
> from some outside factor.  You sure your shell script isn't
> misbehaving and corrupting the reported return code?

Good to know.  The RC might be from ssh (the actual command being "ssh 
postgres@HOSTNAME pg_dump ...".

-- 
Angular momentum makes the world go 'round.

Ron <ronljohnsonjr@gmail.com> writes:

> On 9/22/22 09:34, Tom Lane wrote:
>
>> Ron <ronljohnsonjr@gmail.com> writes:
>>> v9.6.2.23
>>> I always run pg_dump with the --verbose option, and the bash script echos
>>> $?, so there's a 25,664 line log file for me to examine which explicitly
>>> shows the return code.
>> A quick search of the 9.6 pg_dump source code shows that it only
>> ever calls exit() with values 0 or 1.  255 would have to be coming
>> from some outside factor.  You sure your shell script isn't
>> misbehaving and corrupting the reported return code?
>
> Good to know.  The RC might be from ssh (the actual command being "ssh
> postgres@HOSTNAME pg_dump ...".


FWIW, here's what my ssh man page says, and a contrived example.

So it seems your ssh call itself borked for some reason perhaps.

EXIT STATUS
     ssh exits with the exit status of the remote command or with 255 if an
     error occurred.


tmp$ ssh localhost false; echo $?
1
tmp$ ssh no-such-host true; echo $?
ssh: Could not resolve hostname no-such-host: nodename nor servname provided, or not known
255
tmp$