Created attachment 221492 [details] unified patch Maybe it's just me, as I have a couple of nodes which are a little memory challenged and at times die in OutOfMemory errors… but seems a nice idea to auto-restart Cassandra if/when it dies anyways. PS: this is against databases/cassandra3 but the same patch could be useful in databases/cassandra4 as well.
^Triage: [tags] in issue Titles are deprecated. ^Triage: Simplifying title
The second '-f' flag in the command_args is to put cassandra into the foregroung? The first '-f' to daemon redirects all IO to /dev/null, so what is the overall effect?
Yes, Cassandra needs to remain in the foreground, as without `-f` parameter it starts a (separated) daemon process and the launch program closes immediately (and thus `daemon` would think it died and re-start it immediately).
The first `-f` does the same that `>/dev/null 2>/dev/null` does in the current version.
Already committed by bug #253312 (in r564833).