We're getting the following error when initiating a deployment in our new RA environment: Failed to distribute artifact [artifact_store/releaseId_<releaseId>/...] to execution server [es_<ExecutionServername>].
CA Release Automation 5.x and 6.x
The job's artifact package was being delivered to an agent reporting to one execution server from a retrieval agent that reports to a different execution server. Based on this finding we looked and confirmed that the two execution servers weren't configured to communicate with each other (which is necessary to get a file/artifact across).
There are a few solutions to this kind of problem:
- Configure the execution servers to communicate with each other. Instructions for establishing this connection can be found here: Create Sibling Connections Between Execution Servers
- If configuring two remote execution servers to communicate with each other is not possible due to network restrictions then and you have three or more execution servers you can configure one execution to act as a routing execution server by configuring that one connect to all other execution servers. There are some different architectures discussed/described in the Architecture and Implementation Best Practice Guide.
- Duplicate your artifacts - one for each location that has connectivity between the retrieval agent, its execution server, and the agent machines included in the deployment.