Thread

1. archive_command debugging

   postgres@arcict.com — 2023-08-23T15:07:01Z

   Hello,
   
   I was wondering if anyone had any good tips for "debugging" the archive_command in the postgresql.conf.
   
   I want to see what certain settings are when the command is run, like $PATH and id, but I don't want it to be successful either, so that the WAL files remain untouched.
   
   Any tips?
   
   Thanks,
   
   Nick Renders
   
   
   
   

2. Re: archive_command debugging

   Christoph Moench-Tegeder <cmt@burggraben.net> — 2023-08-23T16:03:23Z

   ## Nick Renders (postgres@arcict.com):
   
   > I was wondering if anyone had any good tips for "debugging" the
   > archive_command in the postgresql.conf.
   
   For starters, you get rather noisy logging when that command fails
   (plus all the output from the command itself), so check your postgres
   logfile first. Also, you get some additional debug logging at level
   DEBUG3.
   Then archive_command basically inherits it's environment and user id
   from the archiver process, so you could check that process' proc entries.
   Third, archive_command is executed via system() (after template
   processing), so you're free to do... quite a lot in there (just
   don't mess with the exit status). It might be more convenient to
   use some shell wrapper around your final archive command and have
   that wrapper do all the logging etc. you might want.
   
   And a final tip: do not rely on the environment, always use full
   paths, and if you ever might have whitespace or other "funny
   characters" in some path/file name (or always, just to be safe)
   make sure you get your quoting right.
   
   Regasrds,
   Christoph
   
   -- 
   Spare Space