Oracle MVA

Tales from a Jack of all trades

Archive for September 2009

11G R2 release party

leave a comment »

Yesterday I went to the 11G R2 release party of Oracle the Netherlands. It was great to see the usual suspects of the Dutch DBA crowd. Obviously this was a marketing fling, but Lucas Jellema‘s presentation about Edition Based Redefinition was most interesting. Alex Nuijten is presenting about the subject on planboard, which is yet another reason to visit that seminar (sorry, dutch only).

Written by Jacco H. Landlust

September 24, 2009 at 12:27 pm

Posted in Announcements, RDBMS

BPEL dehydration store and HW contention

with 2 comments

Today I spent some time on fixing performance issues with the BPEL dehydration store. The database is a three (3) node RAC cluster, running RBDMS on ASM BPEL connects to a service name that has one (1) preferred instance and two (2) available instances.

HW contention usually means: trouble with lobs. Just to be sure, I started to check metalink. First of all I discovered I had to set event 44951 to fix bug 6376915 (setting an event without restarting the database means setting events =, next set the event in the spfile). This didn’t clear my HW contention problems though.  What did solve the problems, was moving the lobsegments to a separate tablespace. Here’s the script I used to move the lobsegments (with some extra segment space shrinking):

Read the rest of this entry »

Written by Jacco H. Landlust

September 17, 2009 at 9:26 pm

Posted in BPEL, RAC, RDBMS

Enable archivelog through listener

with one comment

Suddenly I found myself migrating a database from 8.1.7 to on Windows (don’t ask 😉 ). Migration steps were rather easy:

  • Set NLS_LANG to proper value
  • Create export of user (it was only one user)
  • Check privileges and tablespaces
  • Create new database (not in archivelog mode for performance reasons)
  • Create tablespaces
  • Create user and grant privileges
  • Import user
  • Enable archiving and schedule rman backup *ERROR!!!*

All of the sudden I found myself being able to create a database, being member of ORA_DBA privilege group, but not able to connect without the listener. This means that when you shutdown the database (before you can put it in mount phase) your connection is *gone*.

Now what !? I’m no Windows-man. Never have been, never will be.

After a cup of coffee is tried just close the database (alter database close), enable archivelog mode (alter database archivelog) and restart the Windows service. Jeez this is dirty, but it works. Time to have a colleague find out what’s wrong with the domain user I used 😉

Written by Jacco H. Landlust

September 17, 2009 at 7:31 pm

Posted in RDBMS

Suddenly EUS stops working….

with 3 comments

For one of my clients I am assisting on a EUSimplementation with RDBMS and OID all on OEL 4.7. After implementing EUS and enjoying using my personal credentials instead of working as sys or system all worked like a charm. Customer happy, me happy, everybody happy.

After some time all of the sudden EUS stoped functioning for certain databases. Since the number of databases that EUS was not working for is growing, I was called to find out what was going on. I checked the login by setting event 28033:

alter system set events ‘28033 trace name context forever, level 9’;

Next i tried to login and I read the tracefile in $ORACLE_ADMIN/ORACLE_SID/udump:

Oracle Database 10g Enterprise Edition Release – 64bit Production
With the Partitioning, Real Application Clusters, OLAP, Data Mining
and Real Application Testing options
ORACLE_HOME = /u01/app/oracle/product/10.2.0/db_200
System name: Linux
Release: 2.6.9-
Version: #1 SMP Mon Dec 22 02:43:08 EST 2008
Machine: x86_64
Instance name: ORACLE_SID
Redo thread mounted by this instance: 2
Oracle process number: 67
Unix process pid: 8058, image: oracle@SERVER.DOMAIN_NAME

*** ACTION NAME:() 2009-09-16 14:06:03.492
*** MODULE NAME:(sqlplus@SERVER.DOMAIN_NAME (TNS V1-V3)) 2009-09-16 14:06:03.492
*** SERVICE NAME:(ORACLE_SID) 2009-09-16 14:06:03.492
*** SESSION ID:(527.12773) 2009-09-16 14:06:03.492
kzld_discover received ldaptype: OID
kzld found pwd in wallet
KZLD_ERR: Failed to bind to LDAP server. Err=49
KZLD is doing LDAP unbind
KZLD_ERR: found err from kzldini.

Ldap error 49 is invalid credentials, but I know for sure that my credentials are correct! I even use them to logon to the machine that the database is running using OAS4OS (which was somewhat challenging too, since OAS4OS as provided by Oracle is far from enterprise ready).

Knowing my userentry is correct, I checked the database entry in the oid. To my surprise I noticed that pwdexpirationwarned and pwdgraceusetime was set for the database entry (see cn=DB_NAME,cn=OracleContext,dc=domain,dc=acme). This suggests that the passwordpolicy was enforced for databases too, even though the effective subtree was set to cn=users,dc=domain,dc=acme. Simply removing the attributes for all databases in the OID solved the issue for now:

$ORACLE_HOME/ldap/bin/bulkmodify basedn=”cn=DB_NAME,cn=OracleContext,dc=DOMAIN,dc=nl” attribute=”pwdgraceusetime” value=”” replace=true filter=”objectclass=orclDBServer”
$ORACLE_HOME/ldap/bin/bulkmodify basedn=”cn=DB_NAME,cn=OracleContext,dc=DOMAIN,dc=nl” attribute=”pwdexpirationwarned” value=”” replace=true filter=”objectclass=orclDBServer”

Or for the die-hards:

delete from ds_attrstore
 where entryid in ( select entryid
                      from ct_dn
                     where rdn = ‘cn=db_name’ )
   and attrname in (‘pwdgraceusetime’,’pwdexpirationwarned’);

Please keep in mind that Oracle doesn’t support direct querying on the OID.

Let’s find out what Oracle Support is going to give me for a more permanent solution 😉

Written by Jacco H. Landlust

September 16, 2009 at 3:40 pm

11gR2 grid install, just some random things I noticed

leave a comment »

Obviously I am intending to spend my weekend on 11gR2, luckily my girlfriend is at some golf-tournament 😉 Anyway, here’s a preliminary post with some of the things I noticed while installing 11gR2.

Read the rest of this entry »

Written by Jacco H. Landlust

September 5, 2009 at 2:13 pm

Posted in CRS

Hey Joe, what do you know?

leave a comment »

Oracle RDBMS 11 R2 is out. Open World is not until October 11th, I wonder what the big announcement will be. Maybe Tim Hall is right after all 😉

Written by Jacco H. Landlust

September 2, 2009 at 2:15 pm

Posted in Announcements, RDBMS

OUI crashing

leave a comment »

Today I had some fun trying to install CRS on OEL4 in OracleVM. After setting up the system for the installation (ssh, kernel settings, RPM’s etc) I tried to start the installer (OUI). As soon as I saw the first screen and hit the next button, OUI crashed:

# An unexpected error has been detected by HotSpot Virtual Machine:
#  SIGSEGV (0xb) at pc=0x0000002a959248a4, pid=11755, tid=182894172512
# Java VM: Java HotSpot(TM) 64-Bit Server VM (1.5.0_11-b03 mixed mode)
# Problematic frame:
# V  []
# An error report file with more information is saved as hs_err_pid11755.log
# If you would like to submit a bug report, please visit:

The OUI did not have any logging that pointed me in the correct direction, nog even after adding the -logLevel trace -debug /-J-DTRACING.ENABLED=TRUE -J-DTRACING.LEVEL=2/ options. Some extensive searching pointed me towards the oraInst.loc. oraInst.loc is supposed to give references to the directory and group of the Oracle Inventory, but in my case the oraInst.loc looked like this:

$ cat /etc/oraInst.loc

the inventory_loc parameter is missing. Typo’s, missing or wrong information and wrong casing in the oraInst.loc apparently crash OUI !?

Since no other Oracle products are on the system, I could safely remove the file. Now OUI works like a charm. What’s left is finding out where this file came from. I suspect the base image that was used for installing OEL4 (provisioned from OMS)

Written by Jacco H. Landlust

September 2, 2009 at 1:51 pm

Posted in OUI