Revert "Make sure that after toggling ssl on/off all services are restarted

[jgrassler]

This reverts commit 9d523e7.

Reverting this should fix HA job breakage.

[dirkmueller]

no the problem is not the maintenance mode but the problem is neutron-ha-tool crashing due to not being restarted. we can fix that by deleting the failcounter as I did previously, but that was rejected in the previous review: #3286 (review)

[dirkmueller]

I am working on a patch to fix the rootcause, if you don't have time for that then we can use the crm_failcount removal hack as an interim solution

[dirkmueller]

crowbar/crowbar-openstack#2112 is the fix wip for this

[jgrassler]

So why did this only start to fail after #3286 got merged (that's the reason I opened this PR - to make sure it really isn't the culprit)? Over the weekend before April 15th there were a whole bunch of successful HA job runs, and for cloudsource=susecloud9 they still are green. So somewhere there must be code that didn't make it into susecloud9 and that began to cause trouble from April 15th onward...

[dirkmueller]

@jgrassler very simple, because neutron-ha-tool needs to crash repeatedly several times in a row (like 15 times in a row or so) before pacemaker logs a failure. and in the previous code the crash just happened later, interrupting tempest and other things, than clearly causing a repeatable single fail. now it waits, and you trigger it all the time. before it was only happening when you were particularly unlucky.

[dirkmueller]

also, in most cases the "failcount" check was already executed before the failure was logged (and it does so only once)

[dirkmueller]

please spend your review electrons here: crowbar/crowbar-openstack#2112