GitHub user NicoK opened a pull request: https://github.com/apache/flink/pull/3120
[FLINK-5482] QueryableStateClient does not recover from a failed lookup due to a non-running job This PR checks each cached lookup query whether it is complete and removes any failed lookup from the cache in favour of a retry. An appropriate unit test is added based on existing test code. You can merge this pull request into a Git repository by running: $ git pull https://github.com/NicoK/flink flink-5482 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/flink/pull/3120.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #3120 ---- commit 086aca06674618cb3be962a883efa62c77aa1c66 Author: Nico Kruber <n...@data-artisans.com> Date: 2017-01-12T15:41:30Z [FLINK-5482] share more code in QueryableStateITCase commit a50e155cf2a1e7e04b160d5226f16f017509799e Author: Nico Kruber <n...@data-artisans.com> Date: 2017-01-12T15:48:27Z [FLINK-5482] fix QueryableStateClient not re-issuing a lookup upon failure Any failing lookup, e.g. in case the job has not been started yet, previously remained in the lookup cache and thus future queries did not retry the lookup and failed. This commit changes the lookup caching code so that completed and failed futures are removed from the cache and replaced by new lookups. ---- --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---