Client times out on second invocation

kishore

My client times out the second time I tried to invoke methods on
newly acquired proxies.

Essentially I launch the client on the terminal, with the servers previously lauched. The client successfully acquires the proxies to objects on the servers,
it makes some calls and exits. All goes beautifully. When I relaunch the client,
even though the proxies are acquired, my calls time out. I am using the timeout version of the proxies. I basically have to restart the servers.
My goal is to let the servers run and clients connect and make the calls,
a pretty typical scenario. The client does call the destroy method on the communicators .

I must be missing something. Any help is appreciated.
thanks

mes

Hi,

Common reasons for a client invocation to time out include a blocked server (i.e., a server that is deadlocked, or a server that is busy servicing another request and its thread pool has reached its limit), or a network problem that affects connectivity. It doesn't sound like there's a network problem, so I would focus on the first possibility.

I suggest trying to simplify the problem as much as possible. For example, use only one server at first. I would also run the client with the options --Ice.Trace.Network=3 and --Ice.Trace.Protocol=1, which will display network activity as well as show you the names of the operations the Ice run time is attempting to invoke.

If you are still having trouble, please provide as much information as you can, such as whether you are using Ice services like Glacier2 or IceGrid. You could also supply the tracing output from the options I showed above, and the output from running the servers with the same options.

Take care,
- Mark

kishore

re:Client times out on second invocation

I guess it must have been the thread pool reaching its limit on the server
side. Thanks for pointing me in the right direction. Things are working
as expected.

mes wrote: »

Hi,

Common reasons for a client invocation to time out include a blocked server (i.e., a server that is deadlocked, or a server that is busy servicing another request and its thread pool has reached its limit), or a network problem that affects connectivity. It doesn't sound like there's a network problem, so I would focus on the first possibility.

I suggest trying to simplify the problem as much as possible. For example, use only one server at first. I would also run the client with the options --Ice.Trace.Network=3 and --Ice.Trace.Protocol=1, which will display network activity as well as show you the names of the operations the Ice run time is attempting to invoke.

If you are still having trouble, please provide as much information as you can, such as whether you are using Ice services like Glacier2 or IceGrid. You could also supply the tracing output from the options I showed above, and the output from running the servers with the same options.

Take care,
- Mark