[ZOOKEEPER-3231] Purge task may lost data when the recent snapshots are all invalid

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.5.4, 3.4.13
• Fix Version/s: 3.7.0
• Component/s: server
• Labels: pull-request-available
• Assignee: Ling Mao
• Reporter: Jiafu Jiang
• Created: 29/Dec/18 04:05
• Updated: 10/Sep/21 03:38
• Resolved: 10/Sep/21 03:38

Description

I read the ZooKeeper source code, and I find the purge task use FileTxnSnapLog#findNRecentSnapshots to find snapshots, but the method does not check whether the snapshots are valid.

Consider a worse case, a ZooKeeper server may have many invalid snapshots, and when a purge task begins, it will use the zxid in the last snapshot's name to purge old snapshots and transaction logs, then we may lost data. 

I think we should use FileSnap#findNValidSnapshots(int) instead of FileSnap#findNRecentSnapshots in FileTxnSnapLog#findNRecentSnapshots, but I am not sure.

 

Issue Links

• links to: GitHub Pull Request #1079
• links to: GitHub Pull Request #1239