Re: [Duplicity-talk] GPGError - empty GnuPG log

[Kenneth Loafman]

@ede, if there was an exception we should print that in gpg_failed() prior to the error block from gpg it normally prints.  That's the first time we've needed that!  Yes, a patch would do fine.

...Ken

On Fri, Apr 27, 2018 at 8:13 AM, <address@hidden> wrote:

On 27.04.2018 07:21, Oliver Cole via Duplicity-talk wrote:
> Thanks Edgar,

np Oli,

> On 27/04/2018 02:25, edgar.soldin--- via Duplicity-talk wrote:
>> nope, not without an error stack stack to see where the error was thrown. you can try running in debug verbosity '-v9' (really lot's of output) and "hope" it happens again. maybe this will help pinpointing the issue.
>
> Oh, I had this, but I didn't think it would be useful:
>
> Apr 26 07:58:07 GPG error detail: Traceback (innermost last):
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 1555, in <module>
> Apr 26 07:58:07     with_tempdir(main)
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 1541, in with_tempdir
> Apr 26 07:58:07     fn()
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 1393, in main
> Apr 26 07:58:07     do_backup(action)
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 1511, in do_backup
> Apr 26 07:58:07     full_backup(col_stats)
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 572, in full_backup
> Apr 26 07:58:07     globals.backend)
> Apr 26 07:58:07   File "/usr/bin/duplicity", line 430, in write_multivol
> Apr 26 07:58:07     globals.volsize)
> Apr 26 07:58:07   File "/usr/lib/python2.7/dist-packages/duplicity/gpg.py", line 401, in GPGWriteFile
> Apr 26 07:58:07     file.close()
> Apr 26 07:58:07   File "/usr/lib/python2.7/dist-packages/duplicity/gpg.py", line 286, in close
> Apr 26 07:58:07     self.gpg_failed()
> Apr 26 07:58:07   File "/usr/lib/python2.7/dist-packages/duplicity/gpg.py", line 271, in gpg_failed
> Apr 26 07:58:07     raise GPGError(msg)
> Apr 26 07:58:07  GPGError: GPG Failed, see log below:
> Apr 26 07:58:07 ===== Begin GnuPG log =====
> Apr 26 07:58:07 ===== End GnuPG log =====
>

looks like a corner case, where closing the gpg process fails. the code is not proper at
  https://bazaar.launchpad.net/~duplicity-team/duplicity/0.8-series/view/head:/duplicity/gpg.py#L285
is should actually tell about the exception, but that is silently dropped.

@Ken: would you be willing to accept a patch that also catches the message of the causing exception for the printout?

is there something in your server logs around "Apr 26 07:58:07"?

>> another "hack" could be writing a shell wrapper for gpg, piping all output via tee to the outs _and_ into a log file.
>
> If it happens again I will give that a go.
>
>> also check your syslog/dmesg for oom killing or such during that time.
>
> Nothing in syslog.

not good. i'd recommend redoing the full and see if the error comes up again.

what is your duplicity version?

..ede/duply.net