Skip to Main Content

Berkeley DB Family

Announcement

For appeals, questions and feedback about Oracle Forums, please email oracle-forums-moderators_us@oracle.com. Technical questions should be asked in the appropriate category. Thank you!

Core dumps on AIX (after upgrade from 4.5.20 to 4.6.21)

619227Jan 23 2008 — edited Aug 13 2008
Hi,

we are currently using bdb version 4.5.20 on an AIX 560 with 32 processors for some
massive parallel (threaded) data crunching. Everything works like a charm...

Now i tried to upgrade to the version 4.6.21 and compiled it with:
../dist/configure prefix=/data/database/gom/bdb_4.6.21 exec-prefix=/data/database/gom/bdb_4.6.21 --enable-java
enable-o_direct with-mutex=AIX/_check_lock --enable-diagnostic

These are the same settings (without diagnostic of course) used for the previous release.

If i use now the new library our java app (we are using the java api) keeps crashing in the put method when multiple
threads are waiting to enter a lock (as far as i can observe...)

the last blurb in the console (after core-dump) with diagnostic enabled are:

Module=/data/database/gom/bdb_4.6.21/lib/libdb_java-4.6.so
Module_base_address=DC86F000
Target=2_30_20061012_08722_bHdSMR (AIX 5.3)
CPU=ppc (16 logical CPUs) (0x3ffff0000 RAM)
JVMDUMP006I Processing Dump Event "gpf", detail "" - Please Wait.
[AGBerechnung] ERROR : BDB::assert failure: ../dist/../mp/mp_alloc.c/558: "!SH_CHAIN_HASNEXT(bhp, vc)"
./runApp.sh[43]: 500072 IOT/Abbruch-Trap

Does anybody have an idea what could be wrong?

Btw: i've already tried to use pthreads instead of the AIX/check_lock. The result is the same...

Regards,
Chris
Comments
Locked Post
New comments cannot be posted to this locked post.
Post Details
Locked on Sep 10 2008
Added on Jan 23 2008
31 comments
11,616 views