From d6d3b1c9a3461a9e82814b213b3ce69474049963 Mon Sep 17 00:00:00 2001
From: Joe Rafaniello <jrafanie@redhat.com>
Date: Thu, 17 Aug 2017 10:05:38 -0400
Subject: [PATCH] Remove seemingly unnecessary ignoring of SIGTERM

See the same change in ovirt:
https://github.com/ManageIQ/manageiq-providers-ovirt/pull/80

It was added originally back in 2011 in manageiq:
946657b11a0a65abbcdca9975206d71f7c2b7afc

This was actually copied from the vmware vim monitoring events
code from 2009: 22a911716b40428b6d822f0f1e95aabd022e59c9

```
def monitorEvents
  raise "monitorEvents: no block given" if !block_given?

 trap(:TERM) { $log.info "monitorEvents: ignoring SIGTERM" }
...
```

First of all, trap with logging doesn't work since ruby 2.0+:

```
irb(main):001:0> require 'logger'
=> true
irb(main):002:0> trap(:TERM) { Logger.new("stdout").info
"EventMonitor#start: ignoring SIGTERM" }
=> "DEFAULT"
irb(main):003:0> `kill #{Process.pid}`
log writing failed. can't be called from trap context
=> ""
```

See https://github.com/ManageIQ/manageiq/pull/2386

Second, we need to TERM processes in container land.
---
 .../kubernetes/container_manager/kubernetes_event_monitor.rb     | 1 -
 1 file changed, 1 deletion(-)

diff --git a/app/models/manageiq/providers/kubernetes/container_manager/kubernetes_event_monitor.rb b/app/models/manageiq/providers/kubernetes/container_manager/kubernetes_event_monitor.rb
index 6ea56b4276..853834e552 100644
--- a/app/models/manageiq/providers/kubernetes/container_manager/kubernetes_event_monitor.rb
+++ b/app/models/manageiq/providers/kubernetes/container_manager/kubernetes_event_monitor.rb
@@ -13,7 +13,6 @@ def watcher(version = nil)
   end
 
   def start
-    trap(:TERM) { $kube_log.info('EventMonitor#start: ignoring SIGTERM') }
     @inventory = nil
     @watcher = nil
   end