About

Martin Klier

usn-it.de

New job in 2009 – looking back at A.T.U

After working at A.T.U Auto-Teile-Unger headquarters for over six years, there’s a new challenge calling on January 1st, 2009, and hopefully, it’s for a long time again. I guess it’s a good time to look back: […]

read more

Oracle: Remove scheduler jobs in a loop

If there are too much scheduler jobs in an Oracle database, the CJQ process may die unexpectedly. This has nothing to do with job history, not, it’s the number of jobs known to the system. In my experience, the critical number is somewhere around 32.000 in 10.2 64bit. By the way, that’s the solution for […]

read more

Oracle: How to stop a (MTS) dispatcher process – mad behaviour of ALTER SYSTEM

Manually stopping a dispatcher process is one of the things I have to look up each time, since it’s needed not that often. And, the word “dispatcher” does not appear in the command line! 🙂 alter system shutdown immediate ‘D000’; where “D000” is the dispatcher process name selected from v$dispatchers or gv$dispatchers. (In the latter […]

read more

Attended “Reactive Performance Management” with Craig Shallahamer

Monday and today (Tuesday) I attended a 2-days Oracle University class with Craig Shallahamer from OraPub, taking place in Munich. Honestly, I have had no expectations at all, so there has been nothing to be crestfallen or to be fulfilled. I just was looking forward to the one or other hint how to find the […]

read more

memlock config for Debian Lenny

Hi, today I tried to start Oracle XE with parameters “pre_paged_sga=true” and “lock_sga=true” on my Debian Lenny toybox. But Lenny has had a rather strict and really sticky value for the user’s maximum amount of pinned memory (ulimit -l) value: 32 (kb). First attempt, change /etc/security/limits.conf: oracle – memlock 1073741824 Result: Nothing, “ulimit -l” as […]

read more

Oracle DML Parallel Execution Don’ts

Have you ever waited on a parallelized statement? Parallel execution is said to be fast, efficient, system-exhausting. Far from that! There’s a 10.2.0.4 x86_64 system with 16 cores and over 600MB/s write-IO ability, one statement running, one CPU burning, one developer waiting. For days. Finally, at the end of the week, the admin is involved. […]

read more

Nerd compliments.

If anybody ever considers joining irc.freenode.net #oracle (what I’d recommend for all Oracle interested folks), be prepared for being treated like that: <nayyares> it worked !!! <usn> wait some time, it might crash, or restart all cluster services to see if there is a boot bug <nayyares> ok, i am monitoring usn <Rudemeister> Im also […]

read more

TEMP tablespace too small

Today I got a mail from my grid control: “Tablespace TEMP1 is 97% full”, acting for one of the single instance servers. No problem, you may think, this might happen. It’s an OLAP database, so I wasn’t afraid of big numbers as well. But as you can see, using all this huge temp tablespace for […]

read more

Cleanly removing an agent or host target from Oracle Grid Control 10.2.0.4

Removing a host/agent in Oracle’s Grid Control can be unnerving: There are enough dependencies between the agent, the (cluster-)database, the listener, the host and so to make your deleting action an endless sequence if you are following them. But GC really becomes stubborn if the agent we are about to remove is unavailable: due to […]

read more

Grid Control startup fails, installUser=%s_userName%

Oracle seems to have an uncommon sense of humor. Executing gcstartup after patching the grid control server to 10.2.0.4 (Linux x86) fails with: /bin/su: user %s_userName% does not exist I am not surprised, since this is windows’ environment variable notation, and unknown to my Linux shells. Solution: In the head of /opt/oracle/grid/oms10g/install/unix/scripts/omsstup replace installUser=%s_userName% with  the username […]

read more

Oracle expdp utility i18n – strange

Export> kill_job Möchten Sie diesen Job wirklich stoppen ([yes]/no): yes Möchten Sie diesen Job wirklich stoppen ([yes]/no): yes Möchten Sie diesen Job wirklich stoppen ([yes]/no): yes Möchten Sie diesen Job wirklich stoppen ([yes]/no): yes Möchten Sie diesen Job wirklich stoppen ([yes]/no): yes Möchten Sie diesen Job wirklich stoppen ([yes]/no): YES Möchten Sie diesen Job wirklich […]

read more

Effectice load balancing for a (web?) service in RAC 10gR2

Hi, had to review service creation with DBMS_SERVICE package for a good-feeling load balancing in Oracle RAC 10gR2 several times – now I have to write it down for myself. 🙂 I needed it for a web service using shared server today, but the generic syntax is useful for all other purposes as well. Basic […]

read more

ORA-00600 [kgeade_is_0]

Have a look at this nice one (alert log), seen in 10gR2 (10.2.0.4 64bit on Linux): ORA-00600: Interner Fehlercode, Argumente: [kgeade_is_0], [], [], [], [], [], [], [] And from the trace file: Current SQL statement for this session: select tablespace_id, rfno, allocated_space, file_size, file_maxsize, changescn_base, changescn_wrap, flag from GV$FILESPACE_USAGE where inst_id != :inst and […]

read more

Disabling (or enabling) general job execution in Oracle 10g

Hi, disabling and enabling job execution in Oracle DB has two ways: If you are using dbms_jobs, it’s handy to set job_queue_processes to zero to disable the execution of jobs, and set it to a value >0 (maybe 100) to enable. If you are using dbms_scheduler, this parameter does not work for you. You will […]

read more

Find out your character set in Oracle DB

Hi, since ‘show parameter’ does not tell you your database charset, use this statement to see how your characters are stored: select * from nls_database_parameters where parameter=’NLS_CHARACTERSET’; Regards Usn […]

read more

Oracle Database with SGA of or larger than 16GB

Hi, using large SGA’s requires 64bit Oracle RDBMS, of course. But furthermore, some OS settings are to be done. SUSE walks it’s own lane, quite as usual, so I will provide two ways for doing without ORA-27102. […]

read more

Oracle 10gR2 Clusterware on Novell/SuSE SLES10 on multipath devices

Hi RAC-DBAs, installing the Oracle Clusterware CRS 10.2 on SLES10 with its multipath devices can be tricky. Background: The tool formatting the OCR disks, clsfmt.bin, is buggy concerning multipath / device mapper devices and you will see root.sh fail and clsfmt spit out errors like: clsfmt: Failure trying to resize OCR file from … bytes […]

read more

Lecture “Highly available Communications Server” (Oracle Database) at DOAG conference 2007

Hi, I’ll give a lecture about high availability with Oracle RAC and Dataguard at DOAG Conference 2007 (annual conference of the German Oracle User’s Group). The talk will be in German. (“Ein hochverfĂĽgbarer Kommunikationsserver mit RAC, ASM, redundant shared media und DataGuard”) When? Thursday, November 22nd 2007 1300hrs Where? CCN CongressCenter NĂĽrnberg Ost Messezentrum D-90471 […]

read more

Scheduler hangs in Oracle Database 10g

Hi, sometimes the 10g scheduler refuses to start jobs any more. I do not know why at the moment, but as a workaround you can just kill -9 the “coordinator job queue (CJQ0)” process from OS level. The process will be named like “ora_cjq0_SID” in a RAC. PMON does a restart of the coordinator after […]

read more

How to fail over a client transparently in a dataguard switchover/failover scenario

Hi, with using dataguard (Oracle’s hot/cold standby database solution) you have the challenge to make a failover or a roleswitch between the primary and the standby database transparent to the clients. If you don’t, you will have a (partially) loss of service: Clients tnsnames.ora’s will need reconfiguration or the application has to use another TNS […]

read more

Lost, deleted or corrupted a voting disk (quorum) in Oracle RAC 10.2 ?

Hi Folx, whenever you lost, deleted or corrupted a quorum aka voting disk in Oracle Real Application Cluster (RAC) 10.2, you have three options: 1) Reinstall your Clusterware, see Metalink Note 279793.1 2) Replay a backup taken with “dd if=/dev/votingdisk of=/backup/votingfile bs=4k” 3) Have a look at Metalink Note 399482.1 and the RAC Administration and […]

read more

How to move or add a controlfile when ASM is involved

Hi folx, sometimes, you may want to add or move a controlfile. On OS file system, this is no problem. Just “shutdown immediate”, change the CONTROL_FILES init parameter, copy or move the controlfile, and startup again. With ASM involved this is not possible, since the ASM does not allow direct move or copy commands. You […]

read more

Grid Control Agent install without Metalink Access

Hi folx, if you ever need to install Grid Control Agent 10g, there will be no option to avoid configuration of your Metalink Access Data. But you may want to avoid this. The trick is simple: Just decline the License Agreement in the middle of the runInstaller process. The function of agent will be installed, […]

read more