[sos-devel] [sosreport] Wrong exit code on fail

[Bryn M. Reeves]

On Wed, Dec 11, 2019 at 01:15:37PM +0400, Denis Egorenko wrote:
> No changes will be made to system configuration.
> 
> 
>  Setting up archive ...
> 
>  No space left on device while setting up archive
> 
> :~# echo $?
> 0

Looks like this has regressed (we've had a few problems with exit
status propagation down the years). The fix for the exit status
itself is fairly simple:

diff --git a/sos/sosreport.py b/sos/sosreport.py
index c13142c7..0a4c304f 100644
--- a/sos/sosreport.py
+++ b/sos/sosreport.py
@@ -1371,9 +1371,9 @@ class SoSReport(object):
             self.ui_log.error("\nExiting on user cancel")
             self._cleanup()
             self._exit(130)
-        except (SystemExit):
+        except (SystemExit) as e:
             self._cleanup()
-            self._exit(0)
+            sys.exit(e.code)
 
         self._exit(1)

The problem happens because whatever status was set by the error
that first called self._exit() is overwritten by the call that
follows the _cleanup() handler with 0. The fix simply re-uses
the original exit status passed via SystemExit.code.

This may be enough for your use case since you just care about
the overall status but looking at current master there's a bit
more work needed in this area. Right now, for each plugin that
is running at the time of failure you'll see a bunch of log
spew like this:

[plugin:dnf] collecting tail of '/var/log/dnf.librepo.log' due to size limit
Traceback (most recent call last):
  File "/usr/lib64/python2.7/logging/__init__.py", line 892, in emit
    self.flush()
  File "/usr/lib64/python2.7/logging/__init__.py", line 852, in flush
    self.stream.flush()
IOError: [Errno 28] No space left on device
Logged from file __init__.py, line 432

It's cosmetic but obviously not ideal: addressing this will be
a bit more complicated but should be something we can get to
before the next release.

Regards,
Bryn.