Ludovic Courtès <l...@gnu.org> writes: > ng0 <n...@libertad.pw> skribis: > >> Is there a time out in shepherd service start/checks, or is the >> "service could not be started" simply a result of when a service >> encounters a fatal error and quits? > > “Service could not be started” just means that the ‘start’ procedure of > the service returned #f or threw an exception. > > To debug this, you’d have to check the service’s log files, or use > #:log-file to redirect the stdout/stderr to a log file. > > HTH! > > Ludo’. >
Thanks. Observed and tried yesterday: when I turn on logging and make the output more verbose, the log file created is root:root and therefore not accesible by the internal logging mechanisms of the gnunet subsystems which as a consequence throw errors that they can not log to this file. #:user and #:group is now set, and some more changes which I didn't publish so far. There are some gnunet errors, so I have to go all the way back 1.5 - 2 years and see if this was fixed at some commit and have to backport the fix 0.10.1 if these errors are caused by gnunet. But seeing that it's functional without system service it's unlikely that this is the case. -- ♥Ⓐ ng0 | PGP keys and more: https://n0is.noblogs.org/ | http://ng0.chaosnet.org