[389-users] ns-slapd segfault during db2ldif export

[Johan Venter]

Hi all,

Last night I attempted to upgrade my 389 cluster to the latest 1.2.6rc6.

The node that I started with is currently running 1.2.6rc3 and after the 
yum upgrade (running on CentOS, so from epel-testing) I tried to follow 
the instructions at 
http://directory.fedoraproject.org/wiki/Subtree_Rename#warning:_upgrade_from_389_v1.2.6_.28a.3F.2C_rc1_.7E_rc6.29_to_v1.2.6_rc6_or_newer.

The db2ldif export of NetscapeRoot was fine, but the export of userRoot 
resulted in either 1 of 2 outcomes:
  1. ns-slapd segfaulted
  2. ns-slapd sat there doing nothing, using no CPU time at all

I tried multiple times with reboots in between and found that (1) only 
happens the first time, then any attempt to rerun the same db2ldif 
command results in (2).

I know this is not exactly a useful bug report, but I was hoping you 
could point me at what I need to do to report properly - can I just run 
gdb on the ns-slapd process and post the results or are there no debug 
symbols in the binary from the testing packages?

Do I need to setup a VM, compile a debug build from source and see if I 
can replicate?

Cheers,
Johan