[ZOOKEEPER-2355] Ephemeral node is never deleted if follower fails while reading the proposal packet

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 3.4.8, 3.4.9, 3.4.10, 3.5.1, 3.5.2, 3.5.3
• Fix Version/s: 3.4.11, 3.5.4, 3.6.0
• Component/s: quorum, server
• Assignee: Mohammad Arshad
• Reporter: Mohammad Arshad
• Created: 18/Jan/16 14:54
• Updated: 23/Jul/21 13:06
• Resolved: 03/Aug/17 15:58

Description

ZooKeeper ephemeral node is never deleted if follower fail while reading the proposal packet
The scenario is as follows:

1. Configure three node ZooKeeper cluster, lets say nodes are A, B and C, start all, assume A is leader, B and C are follower
2. Connect to any of the server and create ephemeral node /e1
3. Close the session, ephemeral node /e1 will go for deletion
4. While receiving delete proposal make Follower B to fail with SocketTimeoutException. This we need to do to reproduce the scenario otherwise in production environment it happens because of network fault.
5. Remove the fault, just check that faulted Follower is now connected with quorum
6. Connect to any of the server, create the same ephemeral node /e1, created is success.
7. Close the session, ephemeral node /e1 will go for deletion
8. /e1 is not deleted from the faulted Follower B, It should have been deleted as it was again created with another session
9. /e1 is deleted from Leader A and other Follower C

Issue Links

• causes: ZOOKEEPER-3040 flaky test EphemeralNodeDeletionTest
• is related to: KNOX-1599 Upgrade zookeeper to 3.4.13 and curator-test to 2.13.0
• is related to: ZOOKEEPER-2348 Data between leader and followers are not synchronized.
• is related to: CURATOR-409 TestingQuorumPeerMain does not work with Zookeeper 3.5.3-beta
• links to: GitHub Pull Request #112
• links to: GitHub Pull Request #304

Sub-Tasks

• ZOOKEEPER-2355 fix for branch-3.4: Resolved