I'm looking for a way to prevent multiple hosts from issuing simultaneous commands to a Python XMLRPC listener. The listener is responsible for running scripts to perform tasks on that system that would fail if multiple users tried to issue these commands at the same time. Is there a way I can block all incoming requests until the single instance has completed?
-
1are the clients aware of the conflict matrix? should they be intelligent enough to schedule themselves according to this matrix? how do they activities get synchronized etc...jldupont– jldupont2009-10-19 15:19:42 +00:00Commented Oct 19, 2009 at 15:19
-
are you using SimpleXMLRPCServer as I thought this was single threaded...SeriousCallersOnly– SeriousCallersOnly2009-10-19 16:23:39 +00:00Commented Oct 19, 2009 at 16:23
-
I am using the SimpleXMLRPCServer, but it appears I can launch the same command from different hosts simultaneously, which is what i need to preventMattB– MattB2009-10-19 17:18:42 +00:00Commented Oct 19, 2009 at 17:18
-
So on closer inspection it is executing sequentially, scheduling jobs as a first come first serve basis. With a little control logic SimpleXMLRPCServer should do exactly what I need it to, thanks all!MattB– MattB2009-10-19 17:37:14 +00:00Commented Oct 19, 2009 at 17:37
3 Answers
I think python SimpleXMLRPCServer module is what you want. I believe the default behavior of that model is blocking new requests when current request is processing. The default behavior gave me lots of trouble and I changed that behavior by mix in ThreadingMixIn class so that my xmlrpc server could respond multiple requests in the same time.
class RPCThreading(SocketServer.ThreadingMixIn, SimpleXMLRPCServer.SimpleXMLRPCServer):
pass
If I understand your question correctly, SimpleXMLRPCServer is the solution. Just use it directly.
1 Comment
Can you have another communication channel? If yes, then have a "call me back when it is my turn" protocol running between the server and the clients.
In other words, each client would register its intention to issue requests to the server and the said server would "callback" the next-up client when it is ready.