Hello all,
I'm trying to figure out what's going on with
https://fedorahosted.org/cobbler/ticket/530. This defect was
originally reported (almost exactly 11 months ago) before 2.0.2, but
2.0.3.1 still suffers from the problem. As written, cobbler will
*never* successfully call shell-style install/post triggers. Because
subprocess_call() is called with logger=None, all the attribute
references on logger fail.
The calling chain: remote.run_install_triggers() calls
utils.run_triggers() and does not provide the logger= keyword
argument.
There's a fix suggested in the ticket, but that just patches the
problem. This seems to be the right fix, to cobbler/remote.py:
--- remote.py 2010-10-28 21:46:28.000000000 -0400
+++ remote.py.new 2010-10-28 21:46:18.000000000 -0400
@@ -1205,7 +1205,7 @@
# time if reinstalling all of a cluster all at once.
# we can do that at "cobbler check" time.
- utils.run_triggers(self.api, None,
"/var/lib/cobbler/triggers/install/%s/*" % mode,
additional=[objtype,name,ip])
+ utils.run_triggers(self.api, None,
"/var/lib/cobbler/triggers/install/%s/*" % mode,
additional=[objtype,name,ip], logger=self.logger)