Archives for November 2010

Nov
30

Top 5 Performance Tips to Virtualizing Microsoft SQL

1.       Spread the Wealth

Try to keep similar workloads spread out on your vSphere Clusters.  Mix and match high Compute programs with your high disk I/O programs.  This way you have less of a chance of starving your resources. Inside your VM you should still split up database and transaction logs as you would in a physical world. Separating the random I/O patterns(data) and sequential I/O patterns (logs) will help your storage system out.

2.       Adjust the Belt Notches

You can achieve higher network throughput by increasing the transmit coalescing size.  Editing the value of Net.vmxnetThroughputWeight from 0 to 128 will help at the cost of increasing latency.  The setting can be found under the Software settings, advanced settings, then Net. Transmit coalescing is available on the VMXNET2, VMXNET3 and 1000g adapters.

 3.       If You Got, Flaunt It

Use the PVSCI adapters. The PVSCI adapter is more efficient because it allows for batching of I/O requests while the hypervisor is looking for more work. VMware has ran tests and found a 6% percent increase with I/O  throughput over the LSI adapter.  I would also add separate PVSCI controller for your data volumes and log volumes. The PVSCI adapter had a caveat that your workload should be over 2,000 IOPS or you may see a increase in latency. With the release of 4.1 now you can use the adapter under either small or large I/O workload.

 4.       If it’s locked,  it must be good

Any time you can stop paging to disk is good thing. You can prevent Windows from paging out the buffer pool to disk by locking the buffer pool memory .

http://support.microsoft.com/kb/918483

If you are going to do this make the balloon driver doesn’t kick in and defeat everything you are trying to do.  If you make sure all the RAM is reserved you shouldn’t have any problems.

 5.       Go Big or Go Home

Using Large Memory pages can reduce the overhead for your ESX servers.

You can see how to get it going at http://blogs.technet.com/b/sql_server_isv/archive/2010/11/30/trace-flag-834-and-when-to-use-it.aspx

There are lots gotchas so read the article and make sure you have a newer processer that has hardware MMU.

The idea is to have a TOP 10, I will add more points when time permits.

-Dwayne Lessner

Nov
21

VMware View Upgrade Part 2 – Hosts and Agents

Phase 2 of the VMware View Upgrade to View 4.5 was simple for the most part.  Upgraded the hosts  to 4.1 and then took care of the VM tool and agents. This was pretty much the case with a few minor snags. Without a snag or two though it would be an upgrade right?

 The Host upgrade to vSphere 4.1 was really smooth expect for two host.  I had one host that would upgrade because of a vim.fault.NoHost error. It turned out that it was because of a orphaned VM. Once the VM was removed from inventory the upgrade kicked off with no problems.   In another datacentre, one host didn’t take the upgrade because after the reboot a memory problem caused the server to die.  Moved the DIMM’s around and got the server back up and running but I don’t have much faith left in the blade.  For now, HA will let me sleep at night and I made  sure that DRS doesn’t move my CEO’s desktop to that host!

It was great to see vSphere and our Equallogic Storage arrays take advantage of  vStorage APIs for Array integration. You could see the speed difference instantly, though I did do a rescan of all the storage so all the properties would show up properly.  I will turn on Storage IO control shortly. I don’t want change too many pieces at once.

For the agent part of the upgrade we did do a lot of prepping beforehand.  I had another host in our test environment that I copied the 4.1 VMware tools from so our team could do some pre-work.  All the base linked clones had their VMware tools and View agent upgraded prior to the weekend. We also had to update the Wyse TCX Suite to 4.0 to have everything compatible.  We have about 250 Wyse V10-L terminals from the View 3.0 days that we still support, the rest of View Environment are the PCoIP Samsung devices.

As a friendly  reminder make sure to do  these steps in order.   

  1. Upgrade the VM tools
  2. Upgrade the agent
  3. ipconfig /release prior to final snap shot.

The VM tools cam wreck your agent by not allowing PCoIP to work.

We had around 60 desktops from the earlier days when we didn’t run linked clones. The tools were upgrade with Update Manager and we have a great scripting  guys on staff that ran PsExec to upgrade the agents on those desktops.

My team did great job of making sure we covered everthing. Hopefully a bunch of happy users come Monday.

Nov
09

VMware View Upgrade to 4.5

Last weekend we upgraded our VMware View Environment from 4.0.1 to VMware View 4.5. We are running our upgrade in two separate phases. Phase one was to upgrade all of the View connection servers and move our infrastructure to 64 bit. Prior starting our upgrade we had all of our Wyse 10VL upgraded to version 7.0.3 and our Samsung Zero clients to Version 3.2.

Upgrading the View Servers was a pretty easy process. I do wish you didn’t have to shut down the log on services on all the connection servers when upgrading the server components. We had all of the upgrade packages waiting on the servers with a shortcut to shutdown the connection services. This was a process that I wanted to see work with my own eyes.

The next step was to upgrade our vCenter and View Composer to 64 bit hardware. The only comment here is too make sure you read the upgrade guides. There are few parts like moving the RSA keys for View Composer I am sure only VCDX’s would know about.

  We did have some snags during the upgrade with the install of vcenter to 64-bit. The ODBC connection worked but the installer wouldn’t let us upgrade because of  having multiple scheme instances. We found a KB article that talked about the fix but we had to call support. I wasn’t comfortable about deleting some redundant tables. I didn’t what that to come back to haunt me.

Upgrading vCenter Server to a newer version fails with multiple schema errors

I also highly recommend cleaning your house before the upgrade. We had an old connection server that we used for short time for a kiosk project. It had been shut down for months but since it was in the ADAM database it would not allow use to carry out recompose and refresh tasks until it was gone.

Removing a standard (replica) connection server from a cluster of connection servers

The last and final step for phase 1 was to restart all the management agents on the vSphere hosts. View Composer was running reconfiguring tasks for the heck of it. Once this was done our problems were solved.

Phase 2 will we begin soon with upgrading hosts to 4.1 and upgrading all of the tools and agents on the VM’s.