I came across an issue with vCenter backups continually failing on Veeam Backup and Replication 5.0. All the rest of the backups ran smoothly, but vCenter continually failed with an error message stating “VSSControl: Failed to freeze guest, wait timeout” and/or ”VSSFreezer: Failed to prepare guest, wait timeout 900 sec”. As I investigated the issue, I uncovered that it could be one of two issues… Read more >>
With the latest version (5.0) of their vCenter product line, VMware released the new virtual appliance edition of the vCenter server solution. Unlike the traditional vCenter installation that requires an underlying Windows operating system, the vCenter Server Appliance (vCSA) runs on SUSE Enterprise Linux Server 11 and can be easily imported into an existing vSphere environment. This new virtual appliance represents VMware’s direction away from Microsoft Windows based servers to lightweight, pre-configured Linux appliances. While this may seem slightly unnerving to some System Administrators who have grown accustom to the Windows GUI, and the persistent patching required to keep their systems safe and stable; the ability to quickly deploy a reliable vCenter instance will eventually make any Virtualization Administrator’s job much easier.
While the benefits to this model are fairly obvious, there are some limitations with the current release. For starters (and for most this may be a “show-stopper”), the vCSA does not support Microsoft SQL databases. Only an external Oracle instance and an embedded IBM DB2 database is currently supported, with the embedded database being only recommended for small pilot deployments. The vCSA is also not recommended for VMware View environments, as the View Composer component currently requires a Windows Server to operate (View Composer must be installed on a vCenter server). There are other drawbacks as well such as no single-sign on support for Windows session credentials, and no IPv6 support. Yet, it’s my opinion that the two issues mentioned previously must be worked out before most organizations will begin adopt this new technology in their production environments.
There may be some roadblocks with the current release of the vCSA, but as VMware has proved with other technologies, they will eventually iron out the faults in future versions. It is clear though that VMware is committed to providing their customers with the most efficient and reliable technologies, while easing the hassle of system management and deployment. More information regarding the vCenter Server Appliance can be found on the VMware vCenter Server webpage: http://www.vmware.com/products/vcenter-server/overview.html
If you’ve seen, either in person or via the web, any of the recent keynotes delivered by VMware CEO Paul Martiz on the journey to the cloud, you understand that the first step in “the cloud direction” is to get your infrastructure as virtualized as possible. (Check out his keynote from EMC World 2011 here )
From there, many companies move in different directions—and that’s to be expected—but if you’re end goal is to get “to the cloud” where you can truly run “IT as a Service”, then you MUST be able to accurately and, in real time, calculate the costs of operating, provisioning, and otherwise maintaining the virtual infrastructure you and your team worked hard to put in place.
Enter VMware vCenter Chargeback.
Read more >>
Ran into this issue late last week, where a customer with a multi-site design had ESX hosts in Site A show up as disconnected in vCenter. The vCenter server runs in Site B– and they just turned up a new circuit between the sites.
VMware has a pretty good KB article on how to troubleshoot issues with ESX(i) hosts going disconnected in vCenter (kb.vmware.com/kb/1003409). I’ve been down that path before, and had a good idea of what to look for, but nothing seemed to be “wrong”. I was able to ping between host and vCenter (host to vCenter, or vCenter to host), resolve names, verified correct management IPs, restarted managed services, etc.
So after spending about 45 minutes troubleshooting and scratching my head, I ran a few protocol analyzers to see what was actually happening on the wire– and pretty quickly saw what the issue was.
Read more >>