To reproduce: simple case: start broker as root, using default location for storage (/var/rhm) shutdown, change to a 'normal' user and try to restart; get segfault or: create a directory for use as store location, start the broker then shutdown chmod -w for all the files in <store-location>/rhm/dat/ try and restart; get segfault
Caused by aborting a transaction (around recovery) that wasn't actually started. Fixed in rev 1432 of the rhmessaging repo.
Fixed in build.330 on qpid-test, build.41 on qpid-test4.
Fixed in qpidd-0.2-24.el5, qpidc-0.2-24.el5, and rhm-0.2-18.el5