Major Description
This is a weird edge case but I've typo'd jobs enough to notice it a few times.
$ aurora restart test/user/prod/service --shards=235 INFO] Performing rolling restart of job test/user/prod/service (instances: [235]) INFO] Restarting instances: [235] ERROR] Error restarting instances: Not all requested shards are active. INFO] Response from scheduler: INVALID_REQUEST (message: Not all requested shards are active.)
Note that job test/user/prod/service doesn't exist, though the error doesn't call that out. Instead, one is lead to believe that the typo lie merely in the shard id specified
A bare restart on non-existent job for contrast:
$ aurora restart test/user/prod/service INFO] Response from scheduler: INVALID_REQUEST (message: No tasks found for query: TaskQuery(owner:Identity(role:user, user:null), environment:prod, jobName:user-service-production, taskIds:null, statuses:[PREEMPTING, RESTARTING, DRAINING, RUNNING, ASSIGNED, KILLING, STARTING, THROTTLED, PENDING], slaveHost:null, instanceIds:null))