[Pljava-dev] pljava kills my postgres server

Looked in the docs and googled a lot this morning and couldn't find any
info on the problem I'm having.

Running PG 8.1.2 on FreeBSD 6 w/ pljava 1.2.0. Got pljava compiled and
installed correctly. Tried to do install_jar and it dumps core on
postgres with no indication in the log of what it's choking on. Anyone
had any trouble with pljava/postgres/freebsd 6?

I can provide more info later, but not just at the moment. just wanted
to know if anyone else had problems get pljava installed on bsd first...

--

Thanks!

Jon Brisbin
Webmaster
NPC International, Inc.

Hi Jon,
Can you please tell me what Java you're using? If you set the
log_min_messages to DEBUG1 you might get something in the log before the
dump that might help shed some light. A stack-trace from the core file
might help too of course.

Regards,
Thomas Hallgren

Jon Brisbin wrote:
> Looked in the docs and googled a lot this morning and couldn't find any
> info on the problem I'm having.
>
> Running PG 8.1.2 on FreeBSD 6 w/ pljava 1.2.0. Got pljava compiled and
> installed correctly. Tried to do install_jar and it dumps core on
> postgres with no indication in the log of what it's choking on. Anyone
> had any trouble with pljava/postgres/freebsd 6?
>
> I can provide more info later, but not just at the moment. just wanted
> to know if anyone else had problems get pljava installed on bsd first...
>
>

Thomas Hallgren wrote:
> Hi Jon,
> Can you please tell me what Java you're using?
On both FreeBSD boxes, which are the ones I'm having trouble with:

java version "1.5.0-p2"
Java(TM) 2 Runtime Environment, Standard Edition (build
1.5.0-p2-jbrisbin_26_oct_2005_11_01)
Java HotSpot(TM) Server VM (build 1.5.0-p2-jbrisbin_26_oct_2005_11_01,
mixed mode)

On my PowerBook, running postgres 8.1.3, which works just fine and I'm
not having any problems with:

java version "1.5.0_05"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_05-83)
Java HotSpot(TM) Client VM (build 1.5.0_05-48, mixed mode, sharing)

> If you set the log_min_messages to DEBUG1 you might get something in
> the log before the dump that might help shed some light.
Mar 17 14:15:12 development postgres[81573]: [2-1] DEBUG: Not using
integer_datetimes
Mar 17 14:15:12 development postgres[81573]: [3-1] DEBUG: Added JVM
option string "-Xmx256M"
Mar 17 14:15:12 development postgres[81573]: [4-1] DEBUG: Added JVM
option string "-XX:+UseAltSigs"
Mar 17 14:15:12 development postgres[81573]: [5-1] DEBUG: Added JVM
option string "-Djava.class.path=/usr/local/pgsql/lib/pljava.jar"
Mar 17 14:15:12 development postgres[81573]: [6-1] DEBUG: Added JVM
option string "-Dsqlj.defaultconnection=jdbc:default:connection"
Mar 17 14:15:12 development postgres[81573]: [7-1] DEBUG: Added JVM
option string "vfprintf"
Mar 17 14:15:12 development postgres[81573]: [8-1] DEBUG: Added JVM
option string "-Xrs"
Mar 17 14:15:12 development postgres[81573]: [9-1] DEBUG: Creating JavaVM
Mar 17 14:15:17 development kernel: pid 81573 (postgres), uid 70: exited
on signal 10 (core dumped)
Mar 17 14:15:17 development postgres[81565]: [2-1] LOG: server process
(PID 81573) was terminated by signal 10
Mar 17 14:15:17 development postgres[81565]: [3-1] LOG: terminating any
other active server processes
Mar 17 14:15:17 development postgres[81565]: [4-1] LOG: all server
processes terminated; reinitializing
Mar 17 14:15:17 development postgres[81578]: [5-1] LOG: database system
was interrupted at 2006-03-17 14:14:32 CST
Mar 17 14:15:17 development postgres[81578]: [6-1] LOG: checkpoint
record is at 0/4E8843D4
Mar 17 14:15:17 development postgres[81578]: [7-1] LOG: redo record is
at 0/4E8843D4; undo record is at 0/0; shutdown TRUE
Mar 17 14:15:17 development postgres[81578]: [8-1] LOG: next
transaction ID: 83632; next OID: 2541667
Mar 17 14:15:17 development postgres[81578]: [9-1] LOG: next
MultiXactId: 1; next MultiXactOffset: 0
Mar 17 14:15:17 development postgres[81578]: [10-1] LOG: database
system was not properly shut down; automatic recovery in progress
Mar 17 14:15:17 development postgres[81578]: [11-1] LOG: record with
zero length at 0/4E884418
Mar 17 14:15:17 development postgres[81578]: [12-1] LOG: redo is not
required
Mar 17 14:15:17 development postgres[81578]: [13-1] LOG: database
system is ready
Mar 17 14:15:17 development postgres[81578]: [14-1] LOG: transaction ID
wrap limit is 1073823458, limited by database "postgres"
Mar 17 14:15:38 development postgres[81583]: [5-1] LOG: incomplete
startup packet

> A stack-trace from the core file might help too of course.
How do I get one of those?

--

Thanks!

Jon Brisbin
Webmaster
NPC International, Inc.

Jon Brisbin wrote:
>
>> A stack-trace from the core file might help too of course.
> How do I get one of those?
>
The easiest way is to follow this Wiki page that I just added:
http://wiki.tada.se/display/pljava/Debugging+in+C
Once you hit cont in the debugger, you will crash. Hopefully the
debugger will trap that and you can then type the command

bt

That will give you the printout (not from the core though, this will be
from the running process).

Regards,
Thomas Hallgren