We're using 2.0.3 and despite configured 10s timeout for both client (embedded in the same process) and silo, we sometimes see 30 second TimeoutException in logs:
System.TimeoutException: Response did not arrive on time in 00:00:30 for message: Request *cli/e89bccba@8e020588->S10.31.5.13:101:268855415*grn/6EF63849/00000000+133890 #289952: . Target History is: <S10.31.5.13:101:268855415:*grn/6EF63849/00000000+133890:>.
I don't understand how that could happen. We have that set in code:
client.Configure<MessagingOptions>(o => o.ResponseTimeout = TimeSpan.FromSeconds(10));
.....
cluster.Configure<MessagingOptions>(o => o.ResponseTimeout = TimeSpan.FromSeconds(10));
This usually happens after ungraceful shutdown of the silo ...
This is an error and we should fix it.
One problem is that the Options system doesn't support inheritance, but MessagingOptions is exposed publicly and looks like the right type to configure.
Unfortunately, MessagingOptions is abstract and implemented by ClientMessagingOptions and SiloMessagingOptions, so you need to configure both of those for now.
I'll open a PR to at least partially rectify this by applying any MessagingOptions configuration to ClientMessagingOptions and SiloMessagingOptions.
Thanks for clarification. I鈥檒l try with this option types
Most helpful comment
This is an error and we should fix it.
One problem is that the Options system doesn't support inheritance, but
MessagingOptionsis exposed publicly and looks like the right type to configure.Unfortunately,
MessagingOptionsis abstract and implemented byClientMessagingOptionsandSiloMessagingOptions, so you need to configure both of those for now.I'll open a PR to at least partially rectify this by applying any
MessagingOptionsconfiguration toClientMessagingOptionsandSiloMessagingOptions.