Saturday, 29 March 2008

Hmmm, DB2 UDB problem during Quickr 8.1 installation

Haven't cracked it yet, but the installation has failed twice with: -

SQL0970N The system attempted to write to a read-only file. SQLSTATE=55009

when trying to create the WPSDB database: -

CREATE DB wpsdb USING CODESET UTF-8 TERRITORY us PAGESIZE 8192

This is a DB2, rather than Quickr, problem as I see the same error when
I run the command: -

db2sampl

whilst logged in as wpdb2ins: -

db2sampl -verbose

Creating database "SAMPLE"...
Attempt to create the database "SAMPLE" failed.
SQL0970N The system attempted to write to a read-only file.
SQLSTATE=55009

'db2sampl' processing complete.

Permissions for the wpdb2ins user's home directory seem OK: -

ls -al /home

-> drwxr-xr-x 5 wpdb2ins wpdb2iad 4096 Mar 29 08:31 wpdb2ins

ls -al /home/wpdb2ins

-> drwxrwxr-x 3 wpdb2ins wpdb2iad 4096 Mar 29 08:31 wpdb2ins

ls -al /home/wpdb2ins/wpdb2ins

-> drwxrwxr-x 3 wpdb2ins wpdb2iad 4096 Mar 29 08:43 NODE0000

ls -al /home/wpdb2ins/wpdb2ins/NODE0000

-> drwxrwxr-x 2 wpdb2ins wpdb2iad 4096 Mar 29 08:31 sqldbdir

etc.

Will compare these permissions against my working WebSphere Portal 6.0
and 6.1 installations, and see where the difference lies.

More to follow ............

1 comment:

Dave Hay said...

Well, I think I've sorted it out - after a bit more digging, I found the following message in db2diag.log: -

2008-03-29-08.31.49.989077+000 E9877G616 LEVEL: Warning (OS)
PID : 2460 TID : 3086173888 PROC : db2agent (instance) 0
INSTANCE: wpdb2ins NODE : 000
APPHDL : 0-15 APPID: *LOCAL.wpdb2ins.080329083148
AUTHID : WPDB2INS
FUNCTION: DB2 UDB, oper system services, sqlomkdirp, probe:100
MESSAGE : ZRC=0x840F0001=-2079391743=SQLO_ACCD "Access Denied"
DIA8701C Access denied for resource "", operating system return code
was "".
CALLED : OS, -, unspecified_system_function OSERR: EACCES (13)
DATA #1 : File name, 15 bytes
/tmp/SQLDIR.LK0

When I checked, the wpdb2ins user was not able to write to /tmp e.g. the command touch /tmp/foobar failed with a similar read-only error.

Still not sure why, given that I've used this VMware image before ( for WebSphere Portal 6.0 and 6.1 installations ), but c'est la vie.

I'm going to restore the base RHEL image, ensure that everyone has write access to /tmp ( I'm going to cheat using chmod 777 /tmp

Watch this space ...

Note to self - use kubectl to query images in a pod or deployment

In both cases, we use JSON ... For a deployment, we can do this: - kubectl get deployment foobar --namespace snafu --output jsonpath="{...