Deploying Workflow with an off-cluster Redis configuration, where Redis itself is running in cluster mode, does not work.
Symptoms include deis logs working immediately after deploying a new app, but not again after some time. Viewing the deis-logger-xxx pod's logs show 100s of:
2017-06-08T10:00:04.75420277Z 2017/06/08 10:00:04 Error executing pipeline: MOVED 2589 172.20.0.39:6379
The MOVED error means the client needs to reissue the request to the named IP as per the Redis docs.
Please update documentation that Redis Cluster is not supported in the short term and investigate implementation of cluster mode in the longer term.
(FWIW: It's quite easy to boot up a Redis Cluster using AWS ElastiCache.)
Deploying Workflow with an off-cluster Redis configuration, where Redis itself is running in cluster mode, does not work.
Symptoms include
deis logs
working immediately after deploying a new app, but not again after some time. Viewing thedeis-logger-xxx
pod's logs show 100s of:The
MOVED
error means the client needs to reissue the request to the named IP as per the Redis docs.Please update documentation that Redis Cluster is not supported in the short term and investigate implementation of cluster mode in the longer term.
(FWIW: It's quite easy to boot up a Redis Cluster using AWS ElastiCache.)