Am 01.09.2011 18:03, schrieb "Jóhann B. Guðmundsson":
> Timer units along with check units ( having native check units which would 
> connect to service ports/sockets would
> be really nice  )  can monitor those daemons and restart the service if it 
> does not respond ( sometimes daemons end
> up in inconsistent state as in they might still be running but do not respond 
> to request I'm pretty sure systemd
> does not detect that )

this is one of the things i am missing to replace my self written scripts 
because
they rely on "netstat -l" and so a service with a running process which is
a zombie not doing its work is detected

* wait 5 seconds if it comes back (happens often on several sverices)
* if not "/sbin/service  svc stop"; killall processname; /sbin/service svc start

i would really love to have a standard method for such things and since
systemd is well thought from the design i think there should be options for
this in service-files

call them however you want

* listening-port
* listening-interface
* restart=notlisten

port/interface is important because some things are only on local-host or
on a vpn-server you will as example not bind a internal smb on the
vpn-interface


Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
systemd-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/systemd-devel

Reply via email to