mprakash
26th July 2004, 08:31
None of our users able to login!!
Error "Detected database server termination"!!

log.oracle8 file is attched

Markus Schmitz
26th July 2004, 08:47
Hi there,

the information is a bit thin, but it looks like the database driver lost it's connection to the database. All the other errors are consequences from this.

Try the following:

a) Is Oracle up and running?
b) Can you conect via sqlplus?
c) Do you use SQLNet?
d) Did you try to reboot the server? It's windows after all!
e) Think also, wether you did any configuration changes, which migt have caused this?
f) Is any of the disks full?
g) Does the alert.log of oracle say anything?
h) Did a (online) backup fail?

Good luck,

Markus

mprakash
26th July 2004, 09:02
We rebooted the server (it was Unix by the way) and it worked!!!

Markus Schmitz
26th July 2004, 09:22
if it is unix, then you should honestly do a bit of investigation how this happened :-)

Seriously, Unix doe snot have this sporadic behaviour. You might have a kernel parameter set too small and by rebooting you freed ressources. But this just means, the error will happen again. So you better look into it.

Regards

Markus

mprakash
26th July 2004, 09:35
Yes we are facing this problem 4th time in the same server same error.

First time it was the issue of the password and once we reset it worked..

Second time the problem was that of disk full

Third time network problem..

this time rebooting!!!

Please let me know what kernal parameters needs to be checked and what are the idle values. This server is on IBM AIX4.3, Baan application B50c

Markus Schmitz
26th July 2004, 10:00
Unfortunately I can not help your here much. I am more familiar with Baan IV on HP-Ux. Anyway I am sure somebody in this forum posted at some stage the AIX Kernel Tuning Guide. Also check the Oracle requirements for AIX on Metalink.

regards

Markus

Martin
27th July 2004, 09:09
Hi @all

semms it is an error on database level.
to verify, please look to the oracle udump directory and post the related trace file.

martin

mprakash
9th August 2004, 07:15
Hi Martin,

Can you let me know where is Udump directory is located.

We faced this problem again today!!!

Restarting the server is solving our problem!!!

dave_23
9th August 2004, 14:07
Ok, you're losing connection to your DB.. maybe the listener or something is crashing...

This:
2004-07-26[09:19:16(UTC-04:00)]:E:root: Pid 499960 Uid 0 Euid 0 Gid 0 Egid 0 Pset root@fpibaan1:151722
2004-07-26[09:19:16(UTC-04:00)]:E:root: user_type S language 2 user_name root tty locale ISO88591/NULL
2004-07-26[09:19:16(UTC-04:00)]:E:root: Errno 0 bdb_errno 110 (End of file reached)
2004-07-26[09:19:16(UTC-04:00)]:E:root:
2004-07-26[09:19:16(UTC-04:00)]:E:root: dbs_errno = 0Error 522 occurred:
2004-07-26[09:19:16(UTC-04:00)]:E:root: Error ORA-3114 occurred during commit.
2004-07-26[09:19:16(UTC-04:00)]:E:root: ORA-03114: not connected to ORACLE
2004-07-26[09:19:16(UTC-04:00)]:E:root:
2004-07-26[09:19:16(UTC-04:00)]:E:root: Error BDB-522 returned.
2004-07-26[09:19:16(UTC-04:00)]:E:root: Flushed at /view/port.7.1d.03/vobs/tt/servers/ORACLE_2/ora_db.c : #70.
2004-07-26[09:19:16(UTC-04:00)]:E:root: ******* E N D of Error message *******

is nothing to worry about, its your printer daemon, it won't stop you from logging in.

This:

2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: ******* S T A R T of Error message *******
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: Log message called from /view/port.7.1d.03/vobs/tt/lib/al_1/al_sig.c: #173 keyword: CORE DUMPED
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: Pid 8155314 Uid 294 Euid 294 Gid 125 Egid 125 Pset ravi@GEICITRIX:11208
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: user_type N language 2 user_name 01wh4733 tty ote locale ISO88591/NULL
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: Errno 0 bdb_errno 110 (End of file reached)
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: Log_mesg: core_dumped() signal 11 cwd: '/appl/baan5/bse/home/06wh4733'
2004-07-26[09:19:20(UTC-04:00)]:E:01wh4733: ******* E N D of Error message *******

is your users dumping core. This is where your problem is..
do a "strings -a " on the core file that was generated "/appl/baan5/bse/home/06wh4733" and see if you can find which binary is dumping core
and if it gives a reason why.

Also check the udump and bdump Oracle directories, as well as your
unix syslog and messages file.

(Look on metalink for info on udump and bdump, ask your OS vendor for the syslog files if you don't know where they are...)

Additionally, you need to post your $BSE/lib/tabledef6.2 and all of the files in $BSE/lib/defaults/
and the output of the followings commands:
ulimit -a
bshell6.2 -v
uname -a

Dave