To Patch or Not to Patch, that is the Question…

I’ve found myself “whining” in the middle of an Oracle Freelist email conversations in regards to patching-  or the actual need to patch.

As a DBA, I have a love/hate relationship with patching.  I love that if there is a bug in a version of Oracle, I can simply log into Metalink, (or My Oracle Support or MOS, what else are we calling it these days? :)) look up the error, verify with the documentation or with Oracle’s support and apply the patch-  Yay me!

Why am I hatin’ on the patch?  My strong feelings are towards the CPU and PSU patching.  Did I get stood up on a blind date with a patch?  How could I have such strong feelings against patching?  No matter how much the ADHD, compulsive, DBA Mommy in me wants to do what’s best for my databases and I’m being told by the experts that patching regularly is the best thing for my databases, there are almost always “side-effects”.  Bugs that emerge from the patching that requires more patching or even worse, complexities that have no patching available currently or planned.

Explaining to management, developers, analysts and developers why new bugs have popped up after you’ve just taken an outage to address one or more bugs with a patch is not fun…

So….as DBA’s, what do we choose?  Should patching be like poor quality immunizations, resulting in secondary diseases being introduced, requiring more immunizing?  Should the cure be worse than the disease?  What is the right balance and when do we justify the risk? 

I applied my first PSU patch-  yes, that’s right, my FIRST PSU patch.  I’ve been a DBA for quite sometime and yet this is my first time applying the patches called PSU or CPU.  I’ve been a one-off patch girl for so long, but I finally took some time, put aside my feelings and didn’t create another work-around.

I performed extensive research before I decided the Oct2009 PSU patch was the correct decision and had the right combination of patches to address the issues we have been working around for sometime.  The patch did require a one-off patch update from Oracle support that conflicted with the PSU Patch, but other than that, (oh and pretty much writing up the correct set of instructions, since Oracle hasn’t even updated the commands that are required for checking conflicts for patches in the instructions!) it all went pretty smoothly.

One week has gone by since I applied the patch to my first production system.  Since we like to pummel Oracle features into the ground at my place of business, I’ve had to apply two work-arounds to address issues that arose post-patch.  We are starting to prepare for our 11g environments- so this is probably my first and last PSU Patch for 10g- sorry Oracle…

11g patching….Woohoo, can’t wait… 🙂

Print Friendly, PDF & Email

Comments Closed

  • Paul Alsemgeest

    I've been less carefull with patching. I just make the needed backups and let's go!

    And on new environments I always try to patch whatever can be patched before creating the first DB of course…

    But I share your doubts 🙂

  • pp

    Can we apply cpu patch over psu patch without upgrading database.
    Suppose in oct we applied psu patch now in feb can we apply cpu patch on that databaae and what will be side effect?

  • All patching should be tested via the oPatch check but in my experience, this isn’t an option. One off patches must also be verified before applying or rolling back and applying the PSU or CPU patch.