Hello there,
In the latest AgentJS release, you can set a specific replicaTime. Can you please try that and see if it helps?
what should be its value replicaTime
It seems like the issue may be related to congested subnets, and that migrating to a less busy subnet could be a solution — see (1) in Subnets with heavy compute load: what can you do now & next steps
Is there a recommended way to deploy a canister in a specific subnet, e.g., https://dashboard.internetcomputer.org/subnet/4zbus-z2bmt-ilreg-xakz4-6tyre-hsqj4-slb4g-zjwqo-snjcc-iqphi-3qe?
dfx canister create takes a --subnet flag. You should be able to pick 4zbus with that
1 Like
Thank you! Are there any drawbacks to using a less popular subnet such as 4zbus?
Relatedly, is there any way to cancel an update call if it’s not executed within 10 seconds, as to avoid hanging state changes?
Once an update call has been submitted, you cannot unsubmit it from the protocol’s point of view. The expiry time guarantees that a call’s execution will not start if the subnet time exceed the expiry. If you submit calls with local time + 10s, then you get the intended behavior. However, note that 10s is rather short for a heavily loaded subnet.
2 Likes
In most cases you will be happier on a less busy subnet. Then only time that I can see it being a drawback is if you make a lot of calls to canisters on a specific busy subnet. Calls between subnets take substantially longer (at least 4 rounds extra) than if you call a canister on your own subnet
1 Like