Le jeudi 20 septembre 2012 à 14:05 +0200, Stefan Botter a écrit :
Hi Frederic,
On Thursday, September 20, 2012 01:45:14 PM Frederic Crozat wrote:
Was a bug report opened for this issue (probably better to have it opened upstream) ?
No,
From what I understand, you are complaining on the fact that service status isn't relying on a "customized" logic, as it was possible with initscripts, but on "hard-coded" logic, based on PID or cgroup existence, right ?
I believe so. But to be clear: I do not care, if the init process is sysv-init or systemd, for the cluster software the init-scripts are sufficient - as long as they work. Far up the thread someone mentioned to get rid of the init script altogether, and that would break the actual functionality. Of course one can create custom resource scripts, but for ease of use the init-scripts worked out of the box.
Then implementing ExecStatus support in systemd would allow support this
feature.
--
Frederic Crozat