[AMQ-3806] Partial index updates can lead to bogus recovery for the vmcursor of a Q on startup

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 5.5.1
• Fix Version/s: 5.6.0
• Component/s: Broker, Message Store
• Assignee: Gary Tully
• Reporter: Gary Tully
• Created: 13/Apr/12 15:36
• Updated: 13/Apr/12 15:45
• Resolved: 13/Apr/12 15:45

Description

seeing

 INFO | Using Persistence Adapter: org.apache.activemq.store.kahadb.KahaDBStore@49f10a67
INFO | KahaDB is version 4
INFO | Recovering from the journal ...
INFO | Recovery replayed 1 operations from the journal in 0.063 seconds.
INFO | ActiveMQ 5.5.1 JMS Message Broker (..) is starting
INFO | For help or more information please see: http://activemq.apache.org/
INFO | cursor for queue://JMS/XXXXX has recovered 10000 messages. 2147483647% complete
INFO | cursor for queue://JMS/XXXXX has recovered 20000 messages. 2147483647% complete

The crazy % is the result of recovering on a store that has 0 messages. The orderIndex seems to still allow iteration (possibly part of a partial index update) and can result in recovering for ever.