VDI IOPs simulation

May 9th, 2012

One of the things I am working on at the moment is a VDI design based on a smallish branch office architecture. If you are familiar with VDI, you will know that one of the key metrics is the number of IOPs that the back-end storage needs to support. Historically, VDI projects underestimated this figure or the storage was unable to deliver what was required.

What would be useful is an I/O simulator, a disk subsystem on which you could tweak the IOPs performance to see what effect it had on guests. This idea has been looked at very successfully by Jim Moyle with the Atlantis ILIO product and he has written some great papers on the results.

But I wondered if a basic form of simulation could be done using VMware’s storage IO control (SIOC) or one of the storage appliances available from the virtual appliance marketplace (StorMagic and FalconStor are two options on the front page of the VA marketplace).

It would be worth investigating these options which may pay great dividends in a VDI design exercise. The Rolls-Royce solution would be a complete simulation of every layer in the VDI stack, of course. Network, broker, server and storage.

Rescue Me

May 2nd, 2012

Thanks to the Ubuntu boot recovery program I was able to get back my bootable Ubuntu fairly easily. Although I had to cut a CD to do it as my original USB Ubuntu was 32 bit and my installed system 64 bit. The recovery tool would not repair a 64 bit system when running from 32. Booting the recovery CD normally did not work either, but safe mode did.

At least as an added bonus it detected and added a menu item for OpenIndiana although it detected it as Xen.

OpenIndiana is the Default

April 28th, 2012

Well, that was easy. My server now boots into OpenIndiana by default. The look and layout is very similar to Ubuntu. One slight gripe is that the OpenIndiana installer has clobbered the original grub, which gave me the choice of booting Ubuntu from another partition. Now I will have to go and reconfigure it to give me dual boot back.

I half expected that to happen since it is typical behaviour of any O/S installed to put it’s own boot loader down. It would be nice however if the installer warned it was going to to this.

It’s nice to be able to have a platform to play with ZFS etc. especially since I haven’t used Solaris (or even Unix) much recently. I have been very happily implementing private clouds, proofing AD RMS, streaming Windows applications with Numecent Jukebox and designing VDI solutions. First things first though…Grub documentation.

OpenIndiana Again

April 28th, 2012

Having a relatively free evening, I am installing OpenIndiana on a partition of my lab Shuttle. The browser I am using to type this in is running from the install DVD. So far it has been a piece of cake and I have just noticed it has finished and is asking for a reboot. Back soon…

IPS for OpenIndiana and Solaris 11 – UKOUG March

March 27th, 2012

I went to the March LOSUG / UKOUG where Chris Ridd gave a (long) talk about IPS – Image Packagaing System for OpenIndiana and Solaris 11. I have not used IPS so I was interested to see how it differed from the old pkgadd system I am familiar with.

He started out by saying there were little or no design docs for IPS but there are some blog postings by Stephen Hahn, Bart Smaalders and Tim Foster about the original ideas. Unfortunately I have not had time to look for or read them.

The comment was telling though as the impression I got at the end was that IPS was a system which lacked exactly that: a design. As a result, as Chris pointed out, the terminology is odd and key functionality is missing. Security, for example. As I understood it, the main conceptual difference is that unlike other packaging systems there is no “package” as such. What you make is a config file which describes the content, including metadata and how to get the files you need from a repository.

Until I play with it (which might be a while!) I’m not going to pass judgement. It was too much new information for me to absorb in one sitting.

Esxi 4.1 configuration issues

March 16th, 2012

Have been looking at an ESX 4.1 cluster recently in which one host was being a bit truculent. After a BIOS upgrade the host seemed to be behaving itself so after a couple of weeks of stress testing with no issues I added it back into the cluster.

However, nothing would migrate to it and I eventually tracked down three separate issues (note to self):

  1. Moving hosts between clusters can confuse the standard switch configuration. You really need to re-create the standard switches each time you add to a cluster.
  2. An I/P address was wrong (typo)
  3. The hash of the NFS datastore UUID was different. On the other hosts in the cluster the datastore had been added with I/P address and on this one it had been added with NFS server name. ESX essentially thought they were different datastores as 4.1 uses a crude hash to create a UUID (I would have though it should pick up a UUID from the datastore itself).

Caveat implementor.

LOSUG – UKOUG February

February 20th, 2012

I missed a few months of LOSUG due to one thing and another so was looking forward to Andrew Watkin’s talk on ‘Solaris 11 Automated Installer Walkthrough’. As usual, an engaging and interesting talk! I hadn’t even realised that Jumpstart was dead, as it quickly transpired in the talk. Jumpstart, Jet, building servers in general was my ‘specialist subject’ for a large part of my Unix sysadmin career. It made me realise how out-of-date my Solaris skills were becoming, not having used it on a daily basis for, well a few years. The last major piece of Solaris work I did was the build engineering for a Solaris 10 rollout and that was all done with Jet and flars.

This new ‘AI’ system in Solaris 11 does seem like work still in progress, although it has been around since 0609. ‘What about images?’, I asked, only to find that no equivalent of flar’s existed in Solaris 11. I really must install some Solaris 11 on my home lab!

P.S. I downloaded the llive DVD and started an install from it. The first stage of the install is transferring cpio files. Perhaps there’s life in the old flar yet!

Learning programming

February 10th, 2012

As my daughter continues with the excellent coding club project, I still fear that programming is being taught “backwards”. By this I mean people seem to be taught a specific language and concepts are introduced as the syntax of the language is learnt. True, I, and most of my peers, learnt programming this way but we were self-taught, typically using BASIC on the school computer, housed in a broom cupboard of the Maths department.

Recently, having looked at various scripting languages: Perl, Python, Ruby to name but a few, I still end up using the same concepts and techniques and just using the syntax of the language at hand. Perhaps if I used one language enough, I would come to learn powerful language specific features. But would that necessarily be a good thing?

Provided you know the basic concepts of programming: variables, iteration, recursion, functions, parameters, you can compute what you want in most cases. Or so I have found. To widen the field, introduce list processing, functional and object oriented programming paradigms.

The basic concepts can be written in almost any language. I would start with these ideas and illustrate examples in various languages.

Internal and External Clouds

January 25th, 2012

Having recently worked on a private cloud implementation, I would like to point out the difference between external and internal/private clouds. They are really quite different things.

Some things are obvious. If you are building an internal cloud, you need to buy hardware and some cloud software solution. If you are looking at an external cloud, you don’t need to worry about any of that. All you are concerned about is getting the service (software as a service) you want from your cloud provider. This is what most people expect when they talk about “cloud”.

To make an internal cloud work, while you will no doubt be providing a service, the biggest challenge is integrating that service into existing infrastructure. For example, organisations have a naming standard, how do you integrate with that? Organisations have asset registers, charging models, directory services etc. It is highly likely that any internal cloud will need to interface with all those existing systems.

Getting cloud software installed and running is the easy bit. Integrating with the other systems is where the challenges are.

Now it’s RoR time

December 30th, 2011

Having scratched the surface of python, I am now starting to look at Ruby on Rails as part of some RedPixie generated work. First impressions: it’s complex! It’s a bit apples and oranges to compare python (or any other scripting language) with Ruby on Rails as the latter is a web application framework and the former is a scripting language. Comparing Ruby itself though is fair and my first impression is that it looks rather “messy” compared to python. I just mean syntactically, and that is based on just a few hours of tutorials. That’s not saying much at all so I will have to comment more when I have learned more.

For rails, I started looking at

railsforzombies.org.

It’s good, but personally I can’t see the wood for the trees. While RoR is in itself large and complex, it is only a fraction of the “web experience” and is built on all (practically relies on) existing web technologies: web servers, page design and layout, MVC architectures, restful design style etc. Understanding all of that is a pre-requisite to starting with RoR.

With that in mind, I found the following links useful:

http://www.xfront.com/REST-Web-Services.html

and various pages under

http://guides.rubyonrails.org/.

Part of me thinks all this technology is piling up like a stack of dirty washing up dishes and if it gets and higher, the whole lot will come crashing down. Functionally it’s all fantastic stuff but increasingly, there’s got to be a consolidation of all these pieces into a technology that doesn’t require you to learn how every plate works from html at the bottom of the pile to ruby-on-rails at the top.