Yellow Bricks

by Duncan Epping

Archives for 2009

VMDK alignment part two…

· ·

Yesterday I posted about the mbrscan and mbralign tool that NetApp created. This triggered Nick Triantos of NetApp to write an excellent detailed technical article on why you need to set the alignment:

When setup routine queries the CHS (Cylinder / Head / Sector) information of the disk from the BIOS, the CHS data doesn’t actually corresponded to anything physical. At large LUN sizes, the S number (sectors per track) is always reported as 63, so partitioning is set at 63 sector starting offset.

The problem is that 63 sectors, corresponds to 32256 bytes or 31.5KB, which does not align with storage array’s logical block size. Physical disk blocks always correspond to 512 bytes in size, but for efficiency and performance reasons, disk arrays use a logical block size unit that is a multiple of the physical block size.

For the complete story visit Nick’s Blog on the NetApp website.

Compare your hosts…

· ·

One of the most promising features in my opinion for the upcoming version of ESX definitely is “Host Profiles”. With host profiles you can ensure that each and every single ESX Host has been installed in the same way. But this feature isn’t available yet, and you would probably like to know if at least all hosts in a Cluster share the same LUN’s and/or Portgroups.

Well, it’s no surprise probably that Hugo Peeters created a script that does exactly that:

This Powershell script generates an overview of any items that are not available to every ESX server in a VMware cluster. These items might prevent your vm’s being vmotioned by DRS or restarted by HA. Pretty serious business, I’d say!

The items involved are:
1. datastores
2. LUNs (important when using Raw Device Mappings)
3. port groups

Hugo exports the output to a nice html file so no more importing to Excel needed or whatever.

Hop over to Hugo and pick up the script. The link is at the bottom of the article!

Check your VM’s alignment…

· ·

I’ve been test driving a script called “mbrscan” for a while now. Today I noticed that this script had already been released a couple of weeks ago. The cool thing about the script is that you can check the alignment of the VM’s filesystem from the outside. The script runs directly on the service console of ESX. NetApp’s Nick Triantos describes it as follows:

There’s a new utility that has been added, called mbrscan. The purpose of mbrscan is to identify whether or not a VM has properly aligned partitions. The script will run successfully under two conditions: a) The VM is powered off b) the VM has VMsnaps. The script will error out if none of the above two conditions are satisfied simply because the VMDK(s) are locked.

This can be very useful to identify possible performance issues within just a matter of seconds. Combine this command with “vmware-cmd” to create and remove snapshots and you’ll have the results for you entire environment within minutes.

Now that you just received your results the real question would be:”How am I going to modify this without destroying the data?”

At this point in time that’s not possible. But I know NetApp is working on a tool called “mbralign”. This tool will make it possible to align the filesystem from outside the VM.

The mbrscan tool is part of the “Unified Host Utilities Kit” which reached version 5. The toolkit isn’t available for everyone, you will need a now.netapp.com/ to be able to download it.

UPDATE: MBRSCAN is a binary not an editable script, and the script version has been floating around for a while now…

Orphaned vmdk’s

· ·

While doing a “mini-healthcheck” at a customer site I noticed a specific Datastore with less than 2% of free diskspace. After a bit of research an orphaned VMDK was found. Orphaned vmdk’s are virtual hard-disks that are not connected to a VM. Probably because they were removed from the inventory without deleting the files.

You can easily find these orphaned vmdk’s via the Service Console:

find -iname “*-flat.vmdk” -mtime +7

For those that don’t like using the Service Console you can also check this with Powershell Ad van Bokhoven created a nice script which he describes as follows:

This script asks the virtual center what the disk are of each VM and puts this into an array. After this, it reads all files on all datastores. If the file is a vmdk file, it will check wheter this file is in the array. If it’s not, you’ve found a orphaned vmd.

I would advise to regularly check your environment on orphaned disks, it can save precious diskspace.

vExpert, who will you be nominating?!?

· ·

I already announced it yesterday, but the Nomination form actually was released today. John wrote a great blog article on vExpert of which the following is an outtake:

If you aren’t active in online places like the VMware Communities or the blogs, or if you aren’t active in your local virtualization community, then this award isn’t for you. However, if you’ve been thinking about getting started, I hope these awards are an inspiration. Being active in the community is a great way to get exposure to technologies that you don’t normally touch in your day job and to keep learning every day. (As the Roman smart dude Seneca put it “We learn by teaching.”)

I think this really captures the essence of what being a blogger / active VMTN member is about “we learn by teaching”. This is definitely what got me started and keeps me going. Writing down your thoughts clears your head and writing a technical article forces you to really understand the topic.

Now a lot of people have asked me who I will nominate. I guess I can come up with at least 20 people that deserve the nomination. I guess most of them are already nominated by now. I’m not going to list the people I will be nominating, but I guess most of you can probably guess who anyway.

Now headover to the vExpert nomination form and start nominating those who deserve this recognition in your opinion.