Yellow Bricks

by Duncan Epping

Storage

New version of the NetApp NFS best practices doc!

· ·

I just noticed that there’s a new version(4.3) of the NetApp on NFS best practices document online. This document has a new best practice in there that I discussed a couple of weeks ago about the “/NFS/LockDisable” and the “prefvmx.consolidateDeleteNFSLocks” settings(Page 14).

As far as I can judge at this point in time the document seems to be also following the VMware guidelines. NetApp suggests running several commands directly on the Service Console to do the desired changes. Although I do agree with their suggested changes, I don’t agree with the order of the procedure.

On page 14 in both “procedures” NetApp suggests that you enter maintenance mode before applying the changes. Which is definitely something I would also suggest. But they also suggest to exit maintenance mode right before you reboot the server. In theory this could lead to VM’s being VMotioned to the host you’re about to reboot. When this happens the VM’s will be killed without any notice, which could lead to all sorts of problems as you can imagine.

So if you’re about to make the changes NetApp suggests, please change the order and do a reboot first and exit maintenance mode when the reboot is completed.

SRM, it’s just too easy

· ·

You’ve probably also noticed a whole bunch of Site Recovery Manager(SRM) related articles popping up with people installing and configuring it in their home lab:

I love these articles because they are prove of the fact that SRM is really easy to set-up. But, and this actually scares me, it might seem a bit too easy. I said “too easy” because implementing a Disaster Recovery solution isn’t about the tools you are using. The tools, which will make your life a lot easier, are not the most important piece of the puzzle. Indeed PUZZLE.

There a whole bunch of SRM projects going on globally where VMware PSO, the department I work for, is assisting. These projects typically have a duration of 3 to 9 months, while it seems that with the ease of VMware Site Recovery Manager this should be a matter of days.

People tend to forget that the most important thing about Distaster Recovery / Business Continuity is the business. You need to know the organisation and IT environment very well before you can even start:

  • SLA’s? –> RPO / RTO?
  • Which services are most important to the business?
  • Which servers are part of the service?
  • In which order need these be started?
  • Which service have the highest priority?
  • Are there any dependencies between services?
  • What about the desktops?

And these are just a couple of questions one should normally have to answer before even going down the SRM road. The fact that SRM is so easy to setup makes it really hard to actually explain to a customer why a BCDR project will take much longer then he expected. And remember that although SRM is a great tool you would still need to create a Disaster Recovery Plan, SRM will be part of the plan but it needs to be in place!

I’m not saying that you should not go down the BCDR / SRM road, but be sure to be prepared. (read this e-book, it’s good and it’s free) Get to know your “business”, and be prepared for a long engagement… cause my experience is that normally people have a hard time answering really obvious questions.

You will talk to a lot of people who don’t have a clue of what the core business services / applications actually are. And the same goes for the sys admins, dependencies? Why would you want to know about that and how would I know?

Do you know which questions to ask, do you know how to get the right answers… This is why BCDR subject matter experts are needed for SRM engagements, so before you start give VMware a call, or your local VAC partner for that matter and make sure you get the best out of the SRM product.

SVMotion and disk space

· ·

I received this question a couple of times and there’s no real definitive answer written anywhere…

“Does storage vmotion require additional disk space on the source volume?”

The answer is: Yes it does. Storage VMotion uses the snapshot technology to release the lock on the source disk. This snapshot is placed on the source volume. So in other words, all changes that take place during a Storage VMotion are written to the delta file. This delta file, can and will grow fast.

So keep this in mind if you need to storage vmotion a VM because the VMFS volume is running out of diskspace… it might run out of diskspace sooner than you think.

ESX in a Box with Shared storage but…

· ·

I was just rebuilding my “ESX in a box” setup. I wanted to install an iSCSI Virtual San Appliance but why should I? Your answer probably is: well because you need shared storage to do a VMotion / HA / DRS etc. Yes you are completely right I do need shared storage to have these capabilities, but there’s no need for an iSCSI VSA or NFS appliance for that matter.

A while ago Bouke G. of Jume wrote a nice blog on how to set up shared storage without a SAN appliance. In short you just create an additional disk(scsi id 1:0) on the first ESX VM. Close down VMware Workstation and edit your .vmx file. I would suggest a copy and paste of the following lines and remove the duplicate lines. (scsi1:0.filename etc)

scsi1.present = “TRUE”
scsi1.virtualDev = “lsilogic”
scsi1.sharedBus = “VIRTUAL”
scsi1:0.present = “TRUE”
scsi1:0.fileName = “D:\Virtual Machines\shared_disk.vmdk”
scsi1:0.mode = “independent-persistent”
scsi1:0.redo = “”

disk.locking = “FALSE”
diskLib.dataCacheMaxSize = “0”
diskLib.dataCacheMaxReadAheadSize = “0”
diskLib.dataCacheMinReadAheadSize = “0”
diskLib.dataCachePageSize = “4096”
diskLib.maxUnsyncedWrites = “0”

Now copy the correct .vmx entries to the second ESX VM’s .vmx file and just boot them up. It’s as simple as that. Yes I know setting up iSCSI isn’t difficult but this will save you precious memory, especially when running this as a demo kit on your laptop!

Dell Recovery CD fails to recover ESXi version 3.5

· ·

I just noticed this new KB article that deals about not being able to upgrade ESXi on a Dell box because of the fact that the virtual media is attached:

Upgrade to ESXi 3.5 Update 2.
If you cannot upgrade to ESXi 3.5 Update 2, use the following workaround:

  1. Connect to the DRAC through ILO, as follows:
    1. Open the Media tab.
    2. Open the Configuration tab.
    3. Deselect the Attach virtual media check box.
  2. Boot the ESXi system from the recovery CD.
To use DRAC virtual media to perform the recovery, follow these steps:
  1. Attach the virtual media
  2. Using the virtual media, boot the machine.
  3. When the recovery CD is fully loaded, disconnect the virtual media and proceed with the recovery.

Which reminded of the nice I/O errors this Dell DRAC virtual media produces when attached. So be sure to detach the virtual media before you actually run ESX(i). Same goes for Fujitsu blades by the way, when a virtual media has been present it also produces these nice I/O errors:

Feb 12 09:16:47 esx1 kernel: SCSI device sdc: 2097151 512-byte hdwr sectors (1074 MB)
Feb 12 09:16:47 esx1 kernel: sdc: I/O error: dev 08:20, sector 0
Feb 12 09:16:47 esx1 kernel: I/O error: dev 08:20, sector 0
Feb 12 09:16:47 esx1 kernel: unable to read partition table

Which isn’t as bad as it seems, it’s just not able to read the partition. For Fujitsu blades the only workaround I’ve seen so far was to completely disable USB before booting.