Archive for October, 2010

On Friday it was very good fortune to meet up with one of my buddies from the North Carolina User Summit – Mr Raymond Overman. We had a good old chinwag – setting the worlds to rights on all manner of things including IT, VMware and Virtualization. It wasn’t all shop-talk. The other good thing was that I managed to get my sticky paws on two of Raymondo’s “Blues Bowls”.

When not going about virtualizing stuff where ever possible, one of Raymond’s passions is wood-turning. Check out this if you like to see some great photos. http://www.overmanshop.com/

This has evolved into a brand new concept – the blue bowl. Again if you want to see pictures http://www.bluesbowls.com/

I think they are totally unique, and original. And they are homage to those early guitars – when people were so dirt poor they had to make their own instruments and entainment. Ray brought two blues bowls with him – a 3-string and 4-string – and I was so impressed with them – I took both!

Anyway, I’m figuring that in about 20-30 years time when Raymond’s design has gone global – I will be one of this first to have the original Overman, hand-turned by the legend himself. As Ray said “From your lips to god’s ear!”

For some reason WordPress keep on bring up this imagesthe wrong way round…. Even though they preview fine on the MBP…

It’s been some time that I posted some music to my site. I’ve been listening to a lot John Lennon recently. Things are slowly picking up for the anniversary of his death – there was a recent drama about him in the period just before he met Yoko… which inspired me to go out and get a compilation of his best numbers with some live and obscure stuff too… Paloma Faith has been on my playlist as well as Plan B… and I’m looking forward to hearing Belle & Sebastian’s new “long player” which is released on the 11th October. I’m going up to Scotland to see the band play at the legendary “Barrowlands” in Glasgow in Dec. I first saw them live there, and I could easily say it was the BEST gig I have ever been to – for both music and such a warm atmosphere…

Anyway, down to brass tacks. This band are called the Jolly Boys. They look like a bunch of OAPs, because well, they probably are – they go way back and are one of the founders of the original Jamacian SKA sound. I loved this reworking of the classic Lou Reed “Perfect Day” song…

This weeks chinwag is with Stu Radnidge of vinternals.com. He’s an Ozzie based in London working for a large financial institution. Like a lot of bloggers he’s quite active on Twitter too… He also finds time to attend the London VMware User Group too…

Chinwagging with Stu was an absolute hoot, and what ever you do you must watch the first 5 mins!

This was recorded a few weeks ago, before the whether turned – and it was quite warm still…

As ever if you want the MP3 version of the chinwag it’s here – but to be honest it’s much easier to subscribe to the MP3 podcast via iTunes which means the podcast will be download when ever I do them. Alternatively, if you would prefer the videos on your iPAD/iPOD/iPhone – you subscribe to the video version of the Chinwags there (beware they are big!). If you don’t use iTunes, here’s the generic RSS Feed link

In the chinwag we discuss:

• Is VMware eating its own dogfood? We talk about limitations with vCenter availability and scalability – and issues with VMware’s internal QA Process…
• What is a cloud. Are they fluffy, cumulous, dark or rainy? Will there be a Cloud 1.0 or is it a process not a technology?
• What does – Removing Complexity from the Orchestration Layer mean in practical terms that an idiot like me can comprehend? And we kind of digress to talk about backup as an example of automation processes in the cloud…
• Then we got quite philosophical about the IT Industry – comparing it to the airline or car industry…
• We talks about vBlock from the VCE Coalition
• Will we see Backup-as-a-service (BaaaaS!) or Monitoring as-a-service – (MaaaaS!) or are we drifting into Everything-as-as-service EaaaaS!?
• What’s you opinion about VMware’s OEM of Novel SUSE and Standard VP on SUSE. I know folks said VMware would become the next Novell, but I didn’t think they meant literally?

Podcast Video Play Now | Play in Popup | Download (11)

This week winner of the comment competition is Craig Waters of AUSTRALIA. Criag is a Virtualisation Architect and Data Centre specialist with over 15 years experience in IT, a self-employed IT consultant providing solutions utilizing VMware virtualisation with storage and networking infrastructure technologies.  He recently became the chairman of the Melbourne VMUG, and is looking forward to meeting other IT Professionals and sharing my IT knowledge and experience. He’s keen to contribute to the future development of SRM after having designed and implemented the product for his current customer (55 VMs, 6 ESX hosts 4Prod & 2DR, 2 x CX4-240′s with MirrorView/S all going over a 1Gbps microwave link <would you believe>!!!).

This is the comment that won the competition.

My number one feature request would be for a drag and drop style arrangement when configuring the VM startup order on a recovery plan, I tell you we have a relatively small number of VMs (100?s) but it’s still a time consuming job clicking those arrows to get the VMs in the right order based on the companies server shutdown/startup procedures.

Another one that would be good is some DR procedure generation tool which once you’ve configured your SRM would spit out the DR operational procedures for each/all VMs based on a protection group / recovery plan. Keeping this documentation up-to-date is time consuming and every time we add a new VM the whole thing needs reviewing (anything to reduce documentation!)

I just upgraded to SRM 4.1 on vCenter 4.1, so another suggestion would be a 64bit version of SRM including ODBC, I’ve ran some test recovery plans and they’re a lot quicker which will get our recovery time down further, I still love seeing peoples faces when I tell them a) we did a full DR live failover/failback in about 5 hours end to end for all applications in the business and b) we test this procedure every 90 days…

Also another quick one would be changing SRM coding from Perl to Powershell? This way a lot more automation could be achieved and powershell developers could put more into the customisation of SRM…

The real challenge with SRM has always been the dependency on storage, as other people have commented here, the storage layer and how you carve up your LUNs / assign VMs can really dictate how SRM functions, when we first sold SRM to management there was no end of debate on how single VMs should be recoverable using SRM so if I have an operational failure I can just failover my single app… it took some time for management to get their head around this and even now they have to be reminded that DR is an all or nothing proposition.

The only problem with storage change suggestions is that it will ultimately effect the storage vendors who VMware rely on to create the SRA integration, so again not sure where this could go… I’ve heard about the replication features in the next release, but what about some help with the storage configuration?

My final rant would be about the ability to change the state of the VM on the recovery side, if you are not fortunate enough to have a stretched VLAN between Data Centres then IP recoding is required, what about some greater control on how VMs startup, like the ability to control services on a VM or change the VM hardware configuration…?

Anyway, that’s all, hope there’s something there for you Mike to pass on to VMware.