vCenter

 View Only
  • 1.  Need P2V Suggestions (DRS, LSI, HAL, Cleanup)

    Posted Mar 31, 2008 05:44 PM

    I have just started using VMWare converter and run through a few mixed results tests doing some Hot Clones. After reading some documentation and some other threads, I have some questions I was hoping I could get some feedback on. I am currently running two ESX 3.5 servers (latest critical updates) in a HR and DRS cluster. I am managing with Virtual Center 2.5. I am using Converter Enterprise as a plug in for VC.

    -I have read in some documentation that using a DRS cluster as a source for a P to V migration is not supported unless the DRS is set to Manaual Automation. Is this just for the initial migration or can you switch your cluster back to automated after the migration?

    -Most servers that we will be migrating will be 4 proc servers. I would like to have the Virtual Machines set to 1 proc (or 2 at the most). Are there any issues changing this? I would imagine if you change to 1 proc that the HAL on the server would have problems.

    -By default, Converter uses the Bus Logic SCSI driver. I ran two tests with this. One, I left it as is and another where I changed the driver to LSI prior to powering on the machine. There is a noticeable improvement using the LSI. Is there anything I need to keep in mind when changing to the LSI driver prior to the first power on?

    -Once I have converted a server I perform some cleanup. So far I am uninstalling any hardware specific software and editing the network settings. Is it necessary to remove hidden instances of old hardware in device manager? Anything else?

    Any suggestions are always appreciated. Thanks.



  • 2.  RE: Need P2V Suggestions (DRS, LSI, HAL, Cleanup)

    Posted Apr 08, 2008 06:29 PM

    I use HP server here so after P2Ving, there are some services I set to Disabled. You'll know which ones because they will fail to start. Other things I do:

    1. Disable Windows Time service and use the VMTools time.

    2. Keep the OS vmdk separate from the data/app vmdk. My template has a single C: drive partition. If I need another partition, I create another VMDK. Aids portability and such.

    3. HAL migrations can be a one-way trip so I have two templates. One was built off a single CPU physical server, the other off a dual CPU server.

    4. I keep my templates in a workgroup, Makes it easier when adding to different domains.

    5. Some apps included in my template need GUIDs wiped out. They will be regenerated on powerup. Otherwise the mgmt consoles get confused. Examples are: HP SIM agents, WSUS, Symantec Antivirus.

    adam