You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Meaning when we make a grpc connection to a server it's actually subject to the system tcp socket timeout of a couple of hours.
Behavior wise this is very bad in Rita where payment_validator currently tries to call the althea payment validation code for every incoming payment, at which point the code will hang for a few hours before moving on since there is no server on the other end of that rpc url (we just moved it) and there is also no shorter timeout set.
While the contact struct defines a user desired timeout it's not actually ever used anywhere in the code.
Meaning when we make a grpc connection to a server it's actually subject to the system tcp socket timeout of a couple of hours.
Behavior wise this is very bad in Rita where payment_validator currently tries to call the althea payment validation code for every incoming payment, at which point the code will hang for a few hours before moving on since there is no server on the other end of that rpc url (we just moved it) and there is also no shorter timeout set.
In order to resolve this issue every grpc connect, request, and send method needs to use the tokio timeout routine. Example in the top comment on this branch https://github.com/althea-net/deep_space/tree/jkilpatr/add_timeouts
The text was updated successfully, but these errors were encountered: