I just ran the tests against a build from head, and we don't see the issue anymore.

Thanks,

Nathanael

On Tue, 2011-02-01 at 12:25 -0800, Howard Chu wrote:

Nathanael Anderson wrote:

...

I thought I had attached the backtrace when I opened this, sorry. Here is the backtrack from 2.4.23.

Compile Options: ./configure --enable-ndb --enable-passwd --enable-modules --enable-slp --enable-crypt --enable-overlays --enable-dynamic --enable-syslog --enable-debug --enable-auditlog --disable-cleartext --enable-wrappers --enable-ppolicy

mysql-cluster-gpl-7.1.9a binary distribution from oracle.

Line 339 in NdbTransaction.cpp didn't refer to a file that I saw. Let me know if anything else is needed.

Thanks, that helped. This is now fixed in HEAD.

...

#0 NdbTransaction::execute (this=0x7fa1ae235659, aTypeOfExec=NdbTransaction::Rollback, abortOption=NdbOperation::DefaultAbortOption, forceSend=0) at NdbTransaction.cpp:339 339 NdbTransaction.cpp: No such file or directory. in NdbTransaction.cpp (gdb) bt #0 NdbTransaction::execute (this=0x7fa1ae235659, aTypeOfExec=NdbTransaction::Rollback, abortOption=NdbOperation::DefaultAbortOption, forceSend=0) at NdbTransaction.cpp:339 #1 0x00000000004f8608 in ndb_back_add () #2 0x000000000043ce34 in fe_op_add () #3 0x000000000043d695 in do_add () #4 0x0000000000435dd9 in ?? () #5 0x000000000043662d in ?? () #6 0x00007fa1b0f43538 in ldap_int_thread_pool_wrapper (xpool=<value optimized out>) at tpool.c:685 #7 0x00007fa1af7729ca in start_thread () from /lib/libpthread.so.0 #8 0x00007fa1ae27a70d in clone () from /lib/libc.so.6 #9 0x0000000000000000 in ?? ()

...

As always, when you get a core dump, post the gdb backtrace of the dump. Use the most recent release version, and make sure you use a non-optimized debug build with all debug symbols intact. All of these instructions are written on the page where you submit new Issues.