vCent Appliance 5.0 and service stopping issue

I am all over this vCent Appliance that Vmware came up with.  The technology of vCent on Linux has been on the works for years now but it is exciting to see Vmware recognizing the use case as well.  If this is the vision of Vmware moving forward then I am fully onboard with it.  The appliance is so easy to deploy, configure, and maintain.  Patching is simple because it does not depend on Windows patches (more on that later).  Although I have not seen nor applied a patch for this appliance I do not foresee any issues with it.  The architecture concept is easy, it is pretty much the same concept as Apple and the Mac OSX or iOS.  Vmware controls the hardware, in this case virtual hardware and Vmware controls the linux OS and all the patches that comes with it.  Given that I know their historical concept with ESXi, which is “lean is better” (paraphrasing) I have no reason to believe that this is the same idea with vCenter Appliance.  For me, less OS application means less patches to release that is why I embrace the concept of “lean is better”.

Another reason I like this linux appliance is that it is not affected by any GPO changes that sometimes are controlled by other admin group.  As much as possible I like applications that are not dependent on Windows.  Prime example, I don’t like Sharepoint that depends on SQL, IIS, AD.  There’s too many variables that can go against you.  There are so many issues that can go wrong that you may or may not have control.  The Windows version of vCent is succeptible to any GPO security changes issue or any issues that arise for being Windows.  Depending on how active is your Windows security admin group or how careless, you could be in serious trouble and for me it is not worth the headache.  In addition, Antivirus issues and host base firewall and HIPS issue.   The more I think about Windows version of vCent the more it solidifies the reason why I chose the Appliance version.

Here’s the other portion of the story.  It was never smooth sailing with this appliance, that is just too good to be true.  There was one major issue before it can join the rank of all fairy tale story and have a happy ending.  This issue is probably due to the nature that it is the infant version 5.0 and we’re one of the few first adopter to use it on the enterprise.  Here are the symptom:

  • The vmware-vpxd service would stop for some reason and the only way to fix it is to login to the VM and restart the service or reboot the machine.  This happens a number of times a day.
  • Active directory integrated user is broken.   Some admins are able to login and some are not.  All the admins defined on vCent used to be able to login at one point but for whatever reason some are not.

In the past the bandaid is to reboot.  However when administrators are not able to login, the bandaid is to rebuild the embedded database.  This was recommend back in around November by a Vmware reps.  This sucks because this is like having a new appliance and you would have to readd the ESXi’s and reconfigure clusters and everything so this is not a viable workaround.  I know that there needed to be a better solution than this.  Fortunately, Vmware release a KB that ultimately fix my issue, I came across it around the Christmas time or early January.  I have been operating on this KB for two months now on 3 different vCent 5 Appliance and it has not crashed since.  Therefore I am safely assuming that this is THE FIX.  Here are the steps.

– Login ESXi locally and open vCenter Appliance VM console.

– Login as root to vCenter Appliance VM console.

– issue “service vmware-vpxd stop” (without quotes) to stop service

– issue “su db2inst1” to change user

– issue “db2 connect to vcdb” to connect to local database

– issue “db2 update db CFG FOR VCDB USING logprimary 16 logsecond 112 logfilsiz 8192“.  This is the meat and potato.  Note: 2 of my vCent was happy with logfilsiz 8192 but the vCent that I use for Vmware View 5 needed 16gig so you might have to play with this digit depending on how big the infrastructure.  And yes, I am using vCent Appliance for View 5 (not using link-clone technology or localmode)

-issue "service vmware-vpxd start" to start back the service or “init 6” to restart the VM  (I believe I restarted the VM)

That is it, I couldn’t ask for a better ending than that..  I am issue-free after applying this KB.  I hope that helps.

Here’s the KB reference:  http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2006812