I have installed DB2 ESE V8.1 on RedHat Linux AS 2.1. This was a partioned Database installation on 2 linux machines. When I tried to create sample database using db2sampl command I am getting the following error.

SQL1229N The current transaction has been rolled back because of a system error. SQLSTATE=40504

I am getting the same error when I use the create database command.

Following entry is found in db2diag.log file

errno = 111

2004-02-05-19.19.12.012169 Instance:db2inst1 Node:000
PID:3089(db2fcmdm) TID:1024 Appid:none
oper system services sqloPdbQuerySocketErrorStatus Probe:100

I am able to start and stop the instance successfully from both the machines by using db2start and db2stop command

Any Solutions??????

Look in the Message Ref vol 2 Appendix A : This is ECONNREFUSED. The description is

The connection has been refused. If you are trying to connect to the database, check that the database manager and TCP/IP protocol support at the server have been started successfully.
If using SOCKS protocol support, also ensure that TCP/IP protocol support at the SOCKS server has been started successfully.

Hope this helps

The communication protocol used is TCP/IP and it is working successfully since db2start and db2stop command is working properly from both the nodes.

Hi ajithnayak,

did you check that rah and db2_all works from the instance owner's command line?

You can check with

$ rah date

and

$ db2_all date

Johann

Both rah & db2_all command is working fine.

Following is the output of both the nodes for this command

egrep: /db2home/.profile: No such file or directory
Date & Time
Machine name: date completed ok

Since the date and time is displayed from both the nodes with the machine name I believe the command is working fine.

Hi ajithnayak,

the missing .profile gets me startled somewhat. Are you sure you got the instance owner's home directory shared across both nodes?

What is the output of db2diag.log when you do a db2start?

Also, please post the output of

$ db2 get dbm cfg | grep -i dftdbpath

Johann

Hi jsander,

I am sure the instance owners home directory is shared across both the nodes.

in dbm cfg valus of dftdbpath is /db2home/db2inst1

db2start and db2stop works fine without any error message in db2diag.log file. output for the command is as below

$ db2stop
02-17-2004 17:30:03 0 0 SQL1064N DB2STOP procesing was successful
02-17-2004 17:30:06 1 0 SQL1064N DB2STOP procesing was successful
SQL1064N DB2STOP procesing was successful

$ db2start
02-17-2004 17:30:15 0 0 SQL1063N DB2START procesing was successful
02-17-2004 17:30:19 1 0 SQL1063N DB2START procesing was successful
SQL1063N DB2START procesing was successful

Hi,

so how do you explain the missing .profile in your post above?

Johann

Hi,

Missing .profile must be something to do with user creation. I used useradd command to create the users but I am able to logoff and logon using instance owner username. So I hope that should not be any problem, right????

Thanks,
Ajith

I am not completely sure if you've tried this, but in a partitioned db env, the database directory should not be created to locate in an nfs directory. You have mentioned in one of your posts that dftdbpath is /db2home/db2inst1. The command ref clearly specifies this. If you have not tried this already, create a filesystem on each node with the same name but local to the node. Then issue the create database command, using the on <path> clause and specifying this directory as the path. Hope this helps.

I tried creating database in a non shared directory using on <path> clause with same directory name on both the machines but the error (sql1229n) still exists.

Hi,

it is probably time to check if your instance would allow to create a database on a single node. If it does, you know it is a partitioning problem. If not, you know it is an instance installation or configuration problem.

In order to find out, follow these steps:

0) db2stop
1) locate your db2nodes.cfg file and store away a backup copy
2) edit db2nodes.cfg and do away with the second node
3) db2start

Now try to create a database.

Johann

Database creation on a single node (by editing db2nodes.cfg) works but not across 2 nodes. Error (sql1229n) still exists.

Hi,

can you check if either node can write to /home/db2inst1?

Try something like this:

$ cd ~
$ touch log
$ rah 'echo date >> ~/log'

Also check if the instance owner has the same user id on both boxes.

Sorry, this starts looking like a tough one.

Johann

Both nodes can write to /db2home/db2inst1. I tried using touch test.log command its working.

Even rah echo date >> /db2home/db2inst1/test.log is also working properly.

User Id for instance owner (db2inst1) is also same on both the nodes.