Merge lp:~pwlars/utah/utah.dialout-group-again into lp:utah

[Paul Larson]

Yet another place where we need to be really sure that syslog is in dialout group, or things go very badly. Also support both upstart and systemd service restart methods

After a bit of pain trying to get utah to tell me where it was failing today with the wily images, I narrowed it down to somewhere in this region of code. At first I thought it was surely the fact that we were using restart rather than systemctl restart, but that didn't fix it completely. For some reason, it seems like we're still not getting the syslog user into the dialout groups yet when it tries to start here. I've cowboyed this onto aldebaran for now, but I wanted to go ahead and land it too, just in case.

Unfortunately, I do still see random failures even with this patch, but it does seem to be doing better at least.

[lp:~pwlars/utah/utah.dialout-group-again]

1065. By Paul Larson on 2015-06-04

use service restart

[Para Siva (psivaa)]

No approved revision specified.

[Para Siva (psivaa)]

No approved revision specified.

[Para Siva (psivaa)]

No approved revision specified.

Unmerged revisions

1065. By Paul Larson on 2015-06-04

use service restart

1064. By Paul Larson on 2015-06-04

Yet another place where we need to be really sure that syslog is in dialout group, or things go very badly. Also support both upstart and systemd service restart methods

=== modified file 'templates/utah-setup.jinja2'
--- templates/utah-setup.jinja2	2013-05-22 16:56:32 +0000
+++ templates/utah-setup.jinja2	2015-06-04 11:57:57 +0000
@@ -18,13 +18,15 @@
 		sleep 2
 		gw=\`ip route | grep default\`
 	done
+EOF
 
 	# rsyslog may start skipping remote logging if it got too many
 	# messages to send before the network was up. This makes it
 	# start afresh and deliver our messages.
-	restart rsyslog
-	sleep 1
-EOF
+	adduser syslog dialout
+	# We need to support both upstart and systemd restart methods
+	service rsyslog restart rsyslog
+	sleep 10
 }
 
 install_package() {