GitHub user tzulitai opened a pull request:
https://github.com/apache/flink/pull/3882
[FLINK-6565] Fail memory-backed state restores with meaningful message if
previous serializer is unavailable
Currently, without eager state registration, if on restore of memory-backed
states (`DefaultOperatorStateBackend`, `HeapKeyedStateBackend`) the previous
state serializer cannot be loaded (perhaps implementation changed or it was
simply removed from classpath), we could only fail the job because there is no
serializer to read previous state.
Prior to this PR, the job was failing correctly, but without a meaningful
message (only an NPE).
This PR adds a more meaningful message to the failure. It also adds tests
for the memory-backed backends that the failure is as expected.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/tzulitai/flink FLINK-6565
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/3882.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 #3882
----
commit 8f032dae560713af7d5d77dd4b85fb367332fa63
Author: Tzu-Li (Gordon) Tai <tzuli...@apache.org>
Date: 2017-05-12T11:11:25Z
[FLINK-6565] Fail memory-backed state restores with meaningful message if
previous serializer is unavailable
----
---
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.
---
