To keep up to speed with all things VMware using your Twitter account start following these accounts:
@vmware
@vmwareblogs
@vmwarecares
@vmwarekb
@vmwarecareers
@vmwareevents
@vmworld
@planetv12n
@vmware_partners
@vcloud
@vmwarevsphere
@vmwareview
@vmw_workstation
@vmwarefusion
@springsource
@zimbra
@vmwarearmy
@vmwarelabs
@vmwarevroom
Now get adding.
There seems to be a mis-conception that the VMware HA (High Availability) component gives application level available near-zero downtime. In a recent document review I found a reference touting it to be better than Microsoft Cluster Services and could replace the MSCS offering. Err, let’s have a look at the differences.
Why VMware HA isn't MSCS
I could go on but the point I’m making here is that VMware HA IS NOT application aware. VMware HA caters for ESX host loss from the network and can, if enabled, check if a virtual machine is still running by checking VMware Tools (heartbeating).
Microsoft Cluster Services IS application aware. It’s aimed at the high end / high service availability applications such as MS Exchange or MS SQL.
If you can’t suffer application downtime choose the MSCS and stick it in VMware although I would highly recommend reading the VMware and MSCS documentation first – there are caveats.
VMware KB Article 1004617 aligns the support statements of VMware ESX versions to Microsoft Cluster Services
(steps down off soap box)
One of my colleagues recently undertook a VMware HA review and stumbled upon a couple issues after collecting the information and preparing the scorecard. So, if you’re like me and have this installed ready on your laptop you’d be advised to apply these patches.
[Note :- this is only applicable to VMware authorised Partners]
Patch 1 : http://ftpsite.vmware.com/download/rlspsrl/patches/vha30-base562-hotfix1
This happens when the vSphere environment contains a vNetwork Distributed Switch with distributed portgroup enabled for either FT or VMotion or management functionality. This configuration results in failure of one of the rules that determines the portgroup sub type (whether the portgroup is FT or VMotion or Management), due to null pointer exception.
Note: If the vDS contains only VM portgroups or VMKernel portgroup but not enabled for FT or VMotion or Management, then analysis will go through fine.
Patch 2: http://ftpsite.vmware.com/download/rlspsrl/patches/vha30-base562-hotfix2
Known issue reported from the field and we are in the process of making an official patch soon. This happens when the vCenter server machine has all VMware products installed on non default install locations. This means vCenter and as well as VMware Tools are not installed under “C:\Program Files\VMware”. This results in vCenter not generating a diagnostic file (when ‘export diagnostics’ command is issued), which results in VHA analysis error. We created a hotfix to handle this issue that lets VHA go ahead with analysis bypassing the missing diagnostic file.
To apply a hotfix follow,
That’s it.
