[AMQ-4729] mKahaDB master slave needs lock when filtered adapter locks are created on the fly

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 5.8.0
• Fix Version/s: 5.9.0
• Component/s: Message Store
• Assignee: Gary Tully
• Reporter: Gary Tully
• Created: 19/Sep/13 20:54
• Updated: 19/Sep/13 21:09
• Resolved: 19/Sep/13 21:09

Description

<mKahaDB directory="../kahadb">
   <filteredPersistenceAdapters>
      <!-- kahaDB per destinations -->
      <filteredKahaDB perDestination="true" >
        <persistenceAdapter>
          <kahaDB journalMaxFileLength="32mb" />
        </persistenceAdapter>
      </filteredKahaDB>
    </filteredPersistenceAdapters>

problem:

1. Starting the slave with an "empty" database, i.e no queues results in the slave trying to start and ultimately failing as the address is in use. There's now lock. Pushing a message to TEST.FOO and then starting the database works OK. The slave comes up and locks and waits.

2. Now, adding a TEST.FOO1 to the master. The master now has the two queues, TEST.FOO and TEST.FOO1. Kill the master and let the slave take over. The slave only sees TEST.FOO. Not TEST.FOO1.