So I spent all day trying to resolve two problems, these are problems that more than one person has with the new Virtual Infrastructure products and I hope to outline possible workarounds and solutions to them. While there are not any clear solutions at this stage, there are things you can do.

Updated 27/07/06

The first problem is the "No performance data appearing" issue that we are seeing in VirtualCenter 2.0 after an upgrade from VirtualCenter 1.x. This problem basically is that if you upgrade VC, there is a chance you will get no performance data from the hosts on the performance tab. There are a numer of reasons that this could be happening, I believe one to be a bug and the others to be environment specific issues that can be resolved.

1. The first reason that this could happen is due to a firewall blocking the communication with virtualcenter and the esx host. Check out the article on this site for which ports to open to resolve this.
2. The second reason is that virtualcenter may be taking it's time upgrading your performance data from the old 1.x format to the new 2.x format. A number of people have reported that the problem magically went away and VMware support suggested that this may have been the case.
3. The third reason is that you could have database issues, while I will not go into depth on this you should check the health of your database. If you don't know how to do that then you need to speak to a dba.
4. The fourth and final reason could be a bug. If you have previously only had a small amount of performance data for a small farm and a couple of weeks has gone by, well you may be experiencing a similar issue to myself.

The only true instant way of resolving the "no performance data appearing" issue is to initialise your virtualcenter database and start over. This is not a perfect option but it does have a 100 percent success rate. I have seen comments on the VMware community forum where people have opened support calls on this issue, but my path will be database initialisation.

Update : Andy Stumph sent me in a tip to help with this problem. Some people have reported that by deleting the performance data from two tables in the virtualcenter database then the problem can be resolved.

The only negative on this one is losing your history but it does not seem so bad. So you need to clear the data out of the VPX_HIST_STAT and the VPX_SAMPLE tables. The first thing to do is stop your vitualcenter services, then clear the tables out. If you are in sql server, fire up the sql query manager. Then select your virtualcenter database from the dropdown at the top, probably is on master at this point. Then run DELETE VPX_HIST_STAT , and DELETE VPX_SAMPLE. Start your virtualcenter services and there is a good chance the problem has been resolved.

This problem is also tied into the low alarm setting bug. If you have upgraded your virtualcenter database you may also notice that your vm's are always in a state of alarm. The alarms appear to be set at really low levels. This cannot be corrected until you fix the performance data bug outlined above.

The second issue that I have been wrangling with is ESX 3 hosts disconnecting from virtualcenter 2.x and these hosts will not reconnect. Previously I wrote an article on some processes to check if this happens, and how to do it. Well I followed my own process and guess what, all appeared well but the problem was still there. While I have not nailed this issue down 100 percent, I am well on my way. At this stage it looks as if the "chkconfig --add ramcheck" may be causing the problem. When I confirm this I will amend my statement.

It seems the reboot after this command is enabled, the hosts started to play up. Removing the ramcheck by using the command "chkconfig --del ramcheck", then deleting the host from virtualcenter and adding it in again seemed to resolve the issue. So it may well help others to try this. I know I am a bit vague on this second one, but I am halfway through my troubleshooting process and recovering from some international travel... Hopefully I can find a way to replicate the issue on demand and pass the information on to VMware in the form of a support call.

If you have had these issues or know a solution then hit the contact page and let me know!