datastore ?

[jackie77]

Hi Everyone,

I have a VMware ESXi 5 installed with one datastore that is 1.3TB. Is it possible to reduce the datastore size by 300GB and create a second datastore?

Thanks,

Jackie

[a_p_]

Unfortunately there's no way to reduce the size of a VMFS partition/datastore. You would have to create a new datastore, copy/migrate the virtual machines and delete the old one. Btw. VMware supports only one datastore on a single LUN/disk!

André

[jackie77]

Thank you for quick reply.

So let me ask you this, I have VMware 5 installed on RAID with four hard drives. I would need to create another RAID to have a seperate datastore for VMware to support two datastores?

Thanks for your help Andre

[a_p_]

Not sure whether I understand this correctly. With a RAID set you may be able to split the RAID set into multiple logical volumes, which will be presented to the ESXi host as separate disks. Would this be a possible solution for you?

André

[jackie77]

You have answered my questions and I think that will fix my issue.

My main problem is (and maybe I should post this has another question) is that I have one datastore with 6 virtual machines running. One of them being a SQL server, the SQL server is very slow in retrieving data. So I read a couple of papers online about it and they suggested that the SQL server should running on a seperate datastore in order for it to work right. Do you know if that is true?

Thanks,

Jackie

[a_p_]

IMO splitting the RAID will not help improving the speed of the virtual machine, since it is still running on the same physical disks. This may be a possible solution in case of a storage system with presenting a separate LUN or with an additional RAID set (additional disks).

May  I ask you which RAID controller you are using and whether it has battery-buffered write cache to operate in write-back mode? On which guest OS does the SQL server run? If it is Windows 2003 there may be a chance to optimize performance by aligning the NTFS partition and/or use another NTFS cluster size (64kB instead of 4kB) as recommended by MS.

André

[jackie77]

I beleive it's a PERC6/I SAS RAID Controller Card. Whether it has battery-buffered write cache to operate in write-back mode, I'm not sure.

The server is Dell PowerEdge 2900 and the guest OS is windows 2008.

[a_p_]

Well, at least the controller has the option to add 256MB BBU (see link below). I usually work with HP servers and there's a HUGE difference in write speed (x10 and more) between write-through (without cache) and write-back mode. I'm sure you will benefit from adding this option, if it's not already present.

Since you are running Windows 2008, I assume the NTFS partition is aligned (starting at 1,024MB)!? This is the default, unless the OS was upgraded.

André

DELL RAID Controllers: http://www.dell.com/content/topics/topic.aspx/global/products/pvaul/topics/en/us/raid_controller?c=u...

[jackie77]

Thanks, according to the link you provide I do have those features, I have the PERC 6/I Integrated / Adapter

How can I check to see if it's enable, in the BIOS?

The Guest OS is Windows 2008 and I beleive it is starting at 1024. I ran wmic, partition get Index, StartingOffset and this is what it returned.

Index      Name      StartingOffset

1          Disk #0, Partition #0 1048576

1          Disk #0, Partition #1 241660067840

My data is on Partition #1, is that the problem? Should it be on Partition #0?

Thanks,

Jackie

[a_p_]

How can I check to see if it's enable, in the BIOS?

Not sure about the Dell systems. You may take a look at "Configuration" -> "Hardware" -> "Health Status" in the vSphere Client to see whether it shows something like "Battery on Controller".

My data is on Partition #1, is that the problem? Should it be on Partition #0?

Having the SQL data on a second partition is a good practice and since the partition start sector is dividavble by 1MB it is properly aligned. So I don't see any issues with this.

André

[jackie77]

Thank you Thank you. Thank you.

While I was waiting for your response, i tried moving the database that was acting slow from the Partition #1 to the Partition #2 and it worked. Before I was having 15 sec delays in pulling data and now it's pulling in 30 ms. Wow, what a relief.

I did check the VMware Config and it shows under Battery that I have:

Disk Drive Bay 3 ROMB Battery 0:Low

Disk Drive Bay 3 ROMB Battery 0:Failed

System Board 1 CMOS Battery 0:Failed

Should i be concerned about the CMOS battery failed, next to all three items the status is Normal.

Thanks Andre

[a_p_]

You are welcome. Great you were able to solve this issue.

About the failed CMOS battery. I would probably replace it. With a failed CMOS battery the system may not come up with the correct time after a shut down which may cause unexpected issues on the host and probably on the VM's too.

André