8.1.14

Release 8.1.14

Bug Fixes

Broker

• Regression in deploying large payloads (#13233)
• No space left on device while receiving snapshot leads to unhealthy followers (#6232)

Misc

• When the deployment of a process definition failed, clients can still start new process instances of the actually non-existing process definition (#13254)
• Segfault on enabling async scheduled task (#13164)
• StackOverflowError when terminating process with a lot of child processes (#8955)

Maintenance

• Avoid usage of actor in gateway job stream (#13037)

Merged Pull Requests

• Delete flaky test CancelProcessInstanceBanTest (#13310)
• refactor: stabilize container based tests (#13278)
• [Backport stable/8.1] deps(parent): bump zeebe-test-container to 3.6.0 (#13264)
• Reset the diskSpaceAvailable flag correctly (#13260)
• Prevent StackOverFlowError during termination of children (#13258)
• Invalidate process cache on deployment rejection (#13256)
• Mark command as processed even on processing error (#13206)
• fix(scheduler): ensure callbacks added after future is completed are executed (#13139)
• fix: fix log info arguments #13137 (#13138)
• test(qa): fix flakiness by awaiting snapshot on restarted broker (#13112)
• Monitor Snyk projects on release (#13069)
• Allow ActorFuture.onComplete to be called by non-actors (#13043)
• test(broker): wait until records are exported before taking snapshot (#13039)