| Author |
Message
|
| tomsal |
 Posted: Sat Feb 01, 2003 1:02 pm Post subject: Error -998 on Solaris |
 |
|
Newbie
Joined: 10 Oct 2002
Posts: 9
|
$ fmcamain -y FMCDES &
[1] 15296
$
$ FMC31020E The server could not be started.
FMC31000E A database error occurred. SQL code: -998.
[1] + Done fmcamain -y FMCDES &
----------------------------------------------------------------
$ tail -f /export/home/db2inst1/sqllib/db2dump/db2diag.log
2003-02-01-15.14.25.140334 Instance:db2inst1 Node:000
PID:15296(fmcamain) Appid:*LOCAL.db2inst1.030201211424
XA DTP Support sqlxa_start_branch Probe:100
DIA4700E AX interface could not be loaded. Transaction not registered.
--------------------------------------------------------------------------- |
|
| Back to top |
|
 |
| clindsey |
| Posted: Sun Feb 02, 2003 7:08 am Post subject: |
|
|
Knight
Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx
|
|
| Back to top |
|
|
| tomsal |
| Posted: Sun Feb 02, 2003 11:59 am Post subject: |
|
|
Newbie
Joined: 10 Oct 2002
Posts: 9
|
Thanks Charlie, but I keep receiving the same error code!!!
I've checked all of the parameters pointed at the URL you posted, and the only difference I have is that when I start the QMGR with the user fmc and the process is running the user is mqm!!!
This is how I checked it:
$ ps -flu fmc |grep amq
8 S fmc 29375 29076 0 53 20 ? 203 ? 14:28:27 pts/1 0
:00 grep amq
$ ps -flu mqm |grep amq
8 S mqm 29369 29366 0 41 20 ? 1749 ? 14:28:13 ? 0
:00 amqzllp0 -mFMCQM ?
8 S mqm 29367 29366 0 44 20 ? 2060 ? 14:28:12 ? 0
:00 /opt/mqm/bin/amqzfuma -m FMCQM
8 S mqm 29370 29366 0 41 20 ? 1503 ? 14:28:16 ? 0
:00 /opt/mqm/bin/amqrrmfa -t2332800 -s2
8 S mqm 29372 29366 0 47 20 ? 2427 ? 14:28:17 ? 0
:00 amqzlaa0 -mFMCQM -fip0
8 S mqm 29368 29366 0 41 20 ? 1080 ? 14:28:12 ? 0
:00 amqhasmx FMCQM /var/mqm
8 S mqm 29366 1 0 41 20 ? 4158 ? 14:28:11 ? 0
:00 amqzxma0 -m FMCQM
The user I've logged in is fmc, and I started the QMGR with this same user.
Hope this help!!!
Thanks in advance. |
|
| Back to top |
|
|
| vennela |
| Posted: Sun Feb 02, 2003 2:12 pm Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
From what you have posted it is very evident that mqm started the MQGR.
Stop the QMGR and start it under fmc.
Everything should work fine.
-------
Venny |
|
| Back to top |
|
|
| tomsal |
| Posted: Sun Feb 02, 2003 2:37 pm Post subject: |
|
|
Newbie
Joined: 10 Oct 2002
Posts: 9
|
Thanks Venny.
The output you saw on the previous post was under the user fmc!!
We've even created the QMGR with this user, and the funniest part is that when we issue the strmqm command, the processes of mqseries are running under mqm!!
I'm logged as fmc and the result has kept the same for all this time.
We are trying everythig at this point, anything could help.
Thanks anyway. |
|
| Back to top |
|
|
| vennela |
| Posted: Sun Feb 02, 2003 4:08 pm Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
Tomsal:
Login as root and try to start the queue manager. If mqm still starts the qmgr then there is something wrong.
Can you post (or PM) the fmcrc and qm.ini files. Do an "id" and confirm that you are fmc before starting the qmgr.
Have you done all the verifications mentioned in the URL posted by charlie. It's pretty comprehensive.
-------
Venny |
|
| Back to top |
|
|
| clindsey |
| Posted: Sun Feb 02, 2003 5:57 pm Post subject: |
|
|
Knight
Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx
|
I just did some testing on solaris and it appears you cannot verify that
fmc started the queue manager by looking at the owner of the mq processes. This is only valid on aix.
First, let me ask, is this a new installation that has never come up or has this been working and you just began getting this error?
Is there more in your db2diag.log? There is not much available on DIA4700E. Can you post a few more lines? Also look in /var/mqm/errors. You should be logging a message in AMQERR01.LOG at queue manager startup. Does it give the xa_open return code?
Charlie |
|
| Back to top |
|
|
| vennela |
| Posted: Sun Feb 02, 2003 6:42 pm Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
Oh ok.
In that case (verification of whether or not fmc has started the qmgr) I have a bunch of other things that I can suggest you.
1. Is DB2 started (db2start)
2. Is the db2 tp_mon_name set to mqmax
3. Did you change the password of the user-id that you have specified as the user accessing the RuntimeDB( or may be it is expired ). In the qm.ini file in the XAResourceManager stanza the password is specified. Verify that the "user" (specified in the open string) can connect to the DB using the password.
For example:
If it says something like
XAResourceManager:
Name=db2inst1 FMCDB
SwitchFile=/usr/lpp/fmc/db2swit/db2swit
XAOpenString=FMCDB, fmc, fmcdbpasswd
then on the db2 command propmt try to connect to FMCDB
db2> connect to FMCDB user fmc using fmcdbpasswd.
This user-id (if it's fmc) should be in db2iadm1 (or whatever the group is)group.
If you have applied patches to DB2 then use fmczutil to rebind the databases.
Also the versions:
If it's MQ 5.3 then there is a new db2swit file that you need to use.
(Forgive my ignorance about Solaris -- I always do it on HP and AIX and I thought it would do the same on Solaris)
-------
Venny |
|
| Back to top |
|
|
| tomsal |
| Posted: Tue Feb 04, 2003 8:33 am Post subject: |
|
|
Newbie
Joined: 10 Oct 2002
Posts: 9
|
Thanks Venny and Charlie.
I tried the test again with root and the trouble was the same.
The versions we are using are:
MQSeries 5.2.1
WorkFlow 3.3.2
Solaris 2.8
DB2 7.2 FixPack 8
MQSeries is not reporting any error on the logs, only the usual start and stop of the QMGR.
I'm running out of time, and we are trying almost everything to start the workflow in Solaris.
If there is anything you could help me would be great.
Thanks in advance.
Later on I'll post the output from the log files. |
|
| Back to top |
|
|
| clindsey |
| Posted: Tue Feb 04, 2003 9:36 am Post subject: |
|
|
Knight
Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx
|
The sql0998N should show up in the db2diag.log. It should also contain a reason code. From a command shell, enter 'db2 ? sql0998'. This will list all the reason codes for -998 and what they mean. This should give the actual reason for your failure.
Charlie |
|
| Back to top |
|
|
|
|
