TMS message publish do not retry on socket timeout
Description
We have seen cases where die to messaging service restart TMS message publish gets IO socket timeout. As we do not retry on any IO exception this can cause failures to publish.
We have seen this causing issues while provisioning where the failing to publish message after cluster is provisioned marks the provisioning as failed.
Release Notes
None
Activity
Show:
Samik GuptaMarch 13, 2024 at 6:13 PM
Discussed with @Albert Shau . Generally it should be safe to republish program state messages hence it is ok to retry on IOExceptions.
Avinash AcharMarch 12, 2024 at 9:18 AM
Need to check with Albert
Pinned fields
Click on the next to a field label to start pinning.
We have seen cases where die to messaging service restart TMS message publish gets IO socket timeout. As we do not retry on any IO exception this can cause failures to publish.
We have seen this causing issues while provisioning where the failing to publish message after cluster is provisioned marks the provisioning as failed.