7.31.2010

libjox10.so (Permission denied (errno:13)) while applying CPU patch on HPUX

Notwithstanding the successful CPU April 2010 patch deployment on development servers, we ran into some sort of troubles while applying the same on our 6 node production RAC servers on HPUX platform. At some stage whilst  installation going on, the following error has been thrown:


The following actions have failed:
WARNING:OUI-67124:Copy failed from '/u00/CPU_APRIL2010_PATCH/9352191/8568398/files/lib/libjox10.so' to '/u00/app/oracle/product/10.2.0/db_1/lib/libjox10.so'...

Previously when we use to apply the CPU or any interim patches, we use to anticipate troubles like,  'Text file busy genclntsh: Failed to remove ..;.  And this time the error seems to be a different one.  Therefore, we stopped the patch installation to make sure no Oracle binaries are running and to investigate why the copy had failed. 

In the the second attempt, the patch installation had failed again with the following error:

File not backed up from /u00/app/oracle/product/10.2.0/db_1/lib/libjox10.so to /u00/app/oracle/product/10.2.0/db_1/.patch_storage/8568398_Jun_11_2009_12_49_36/files/lib/libjox10.so... '/u00/app/oracle/product/10.2.0/db_1/.patch_storage/8568398_Jun_11_2009_12_49_36/files/lib/libjox10.so (Permission denied (errno:13))'

Apparently the problem is more clear this time, indicating permission issue. When we paid a close look at the file in the context we found that writable permission is missing from the file, as show below:


-rxr-xr-x   1 oracle     oinstall   20041776 Jun 11  2009 libjox10.so

As a precautionary step, we first copied the source file and then added write permission to the file, you may also change permission with 755, using the following chmod OS command:

chmod +w libjox10.so

Upon adding the write permission to the file, the subsequent patch installation attempt went smoothly and the patch deployed successfully.

What's next?.. Well, an another 8 node production RAC environment is awaiting for the same patch. Wish me good luck mates.
Happy reading,

Jaffar




6.30.2010

PMP exam preparation course

It has been a while since I have updated my blog. Well, here is a quick update about what I was been doing past 3 weeks. I have attend an in-house PMP exam preparation course 2 hrs a day for 3 weeks duration and it is delivered by our Manager Mr. Ahmed Darwish who is PMP certified and well known PMP instructor here in Middle East.

There were nearly 30 attendees in the course and we thoroughly enjoyed the lecture. I must say, the way he presents, explains real world scenarios and answering difficult questions was quite impeccable. Personally, I have gain thorough knowledge about project management, such as, what needs to be followed for a successful project. If I come across of an opportunity to manage a project, certainly I gonna apply all those golden rules mentioned in the PMP course.

In my next blog entry, I will try to share our group picture which we took at the end of the course.

Happy reading,

6.05.2010

Oracle 11g OCP upgrade

Friends,

A short update about my Oracle certification. On 2nd June 2010, I have successfully completed my Oracle 11g OCP upgrade exam.



Jaffar

5.09.2010

FAQ: Oracle mutexes

I have come across of an interesting FAQ about Oracle mutexes while surfing the net. The following link would take you to the FAQ: Oracle mutexes page where you can find some very good information about the Oracle mutex behavior and it's functionality:

http://www.sap2048.com/faq-oracle-mutexes-933.htm

happy reading,

Jaffar

5.01.2010

crs_stat has been deprecated in 11gR2

While reading whats new and whats deprecated in Oracle 11gR2, I came across of the following:

  • Command crs_stat has been deprecated  in 11gR2, do not use it anymore, to find out all user resource state use, $GRID_HOME/bin/crsctl stat res -t
  •  By default ora.gsd is OFFLINE if there's no 9i database in the cluster, ora.oc4j is OFFLINE in 11.2.0.1 as Database Workload Management(DBWLM) is unavailable.


By the way, in 11gR2 or above, you can use the below command to find out the clusterware process state:

  • $GRID_HOME/bin/crsctl stat res -t -init