Hi,
Just a thought, but are you maybe using DHCP to assign IP addresses to your hosts in the network? And is it possible that a lease expires and the client winds up with a different IP address? That’s about the only way I can think of a “No route to host” exception happening. That error is from the OS, it really means that there is no route to the host…, it’s not something that Servoy could cause if it wasn’t so…
sebster:
Hi,Just a thought, but are you maybe using DHCP to assign IP addresses to your hosts in the network? And is it possible that a lease expires and the client winds up with a different IP address? That’s about the only way I can think of a “No route to host” exception happening. That error is from the OS, it really means that there is no route to the host…, it’s not something that Servoy could cause if it wasn’t so…
Not possible. DHCP lease time is 3 days and if the pc is up at the moment the lease ends it negotiates the same ip address again. The ip can change only if a host is kept offline for more than 3 days and the problem should happen only at boot time (in the early morning) not at random time.
Any other clues? Anyone else suffering for this issue?
ngervasi:
Anyone else suffering for this issue?
yes.
I receintly setup a small solution for a client, who’s connecting
from his home to my server here at my place. we’re both on
high speed connections (dsl for him, cable modem for me).
he often reports servoy ‘hanging’, sometimes 6 to 10 time a
day. his reports coincide with errors showing up in the servoy
admin log file, like these:
2006-02-14 12:15 : Error flushing message buffer to client 231308f7-9d77-11da-b5c9-dd846a04f96c, Connection refused to host: 192.168.0.100; nested exception is: java.net.ConnectException: Connection timed out: connect
and these:
2006-02-13 11:17 : Error flushing message buffer to client 5a5b8a7e-9ca0-11da-b5c9-dd846a04f96c, Exception creating connection to: 192.168.1.35; nested exception is: java.io.InterruptedIOException: SocketPool empty timeout
I tried adjusting the ping delay to 30 seconds as suggested, but
this doesn’t seem to have helped.
any help on this would be appreciated.
rm.
Robert
Check your clients Router settings - recently had problem with new Router set to disconnect at 60mins out of the box.
Also worth trying the Cron getTime method I posted earlier in the thread as this has virtually eliminated problems for my OSX client.
Graham Greensall
Worxinfo Ltd
thanks for the input Graham;
I changed the query validation last nite as suggested earlier,
and the client is reporting no problems at all so far today!
in fact, he says the entire application seems to be running
faster!
thanks again to all who responded.
rm.
This has been a very interesting/useful thread, primarily about keeping a client from being dropped erroneously.
I’m looking for a technique to do the opposite – to automatically disconnect clients which sit inactive for hours on end. So far I’ve seen nothing appropriate on the server admin pages and the Server Docs don’t seem to cover this issue either.
Is this possible or do we need to wait for this feature to be added to a forthcoming release?
Interesting problem;
isn’t there a logout method somewhere? perhaps you could use
a cron scheduler to call the logout method after a period of time
has passed?
just a thought.
rm.