Bladewatch.com

IT Knowledge Exchange

It was only a matter of time before you could get a degree in virtualization. IBM has created a two-year associates degree in green data center management.

The program, developed in conjunction with Metropolitan Community College of Omaha, teaches skills in virtualization and server consolidation, energy efficiency, security and compliance and business issues that arise in the data center. It’s unclear whether those issues include IT in-fighting between networking, systems and storage teams.

What the program does promise is problem-solving in a real-world setting — if your real world is built on IBM hardware and software. The school’s data center lab includes IBM Power Systems servers running AIX, IBM I and Linux.

Until now, there haven’t been viable options for vendor-neutral virtualization certifications, but companies like VMware, Citrix, Microsoft and Red Hat all offer specialization courses. There are generalized data center certifications and degrees, though those are typically on the graduate level.

I was talking with a colleague about this just the other day, I’d love to see us work towards more empowerment in our training within the IT world? Could an IT degree include such things as ITIL? Oncall/certification for virtualization or operating systems? We need to offer the right balance of theory, of best practice and skills in order to empower our next generation of graduates to be ready for their first/next roles.

Share and Enjoy

Reuters

SAN FRANCISCO/TAIPEI (Reuters) – As the PC industry embarks on its fitful road to recovery, many are betting that Taiwan’s Acer and HP will lead the rebound with surging netbook sales and a strong presence in booming Asia.

These two PC makers are expected to gain from consumer demand in China, India and other resilient Asian markets, even as corporate demand stays weak in the global economic slowdown.

Acer, the world’s No. 3 PC brand, can bank on its strength in the fast-growing low-cost netbook segment, while HP’s broad customer base and product mix will stand it in good stead.

“Acer will more likely recover quicker because they’re doing well in the pockets that they’re in… they’re one of the lower cost providers out there that is still respected,” said Louis Miscioscia, research director at tech-focused U.S. firm Brigantine Advisors.

An interesting article highlighting opportunities in China, India and other markets, as always as one market closes down, there are always other opportunities in other regions/sectors, I wonder if this is similar in the server/blade sectors? We’ll have to see.

Share and Enjoy

Sun.com

Sun/Oracle have published a statement illustrating Oracle’s commitment to the SPARC platform, to Solaris and in having even more teams helping with the sale and support of SPARC/Solaris systems. You can see the details in this url. It’s great to see and I’m genuinely looking forward to seeing what new products/solutions arise from this acquisition and at the same time Oracle committing themselves to the SPARC/Solaris platform for those users/customers heavily invested in the platform, a little re-assurance or re-affirmation of your commitment to your core business and your customers never hurt anyone.

Share and Enjoy

AppSense

AppSense, the leading provider of user environment management solutions for the enterprise, announced that AppSense Environment Manager won the Gold award in the Desktop Virtualization category in the Best of VMworld 2009 awards program. Sponsored by TechTarget, the awards were judged by an independent team of experts and editors from SearchServerVirtualization.com that reviewed and evaluated nearly 200 products across eight categories and chose winners according to innovation, value, performance, reliability, and ease of use.

The awards highlight the most innovative technologies at the show, and SearchServerVirtualization.com judges said AppSense Environment Manager “rocked our boat…It offers the most complete user management environment system out there.”

AppSense Environment Manager is the only enterprise solution that enables standardized desktop environments to be fully personalized without the need for cumbersome profiles or scripts. From server-based computing environments through to virtual and physical desktops, AppSense Environment Manager ensures users always receive a consistent, predictable and personalized working experience. Full desktops can now be configured and business rules applied on-demand,enabling compliant, personalized virtual desktops to be quickly delivered tothousands of users. Additionally, AppSense technology is used in conjunction with many third party systems integrators, including CSC, HP, EDS, Dell and IBM.

Well done to AppSense for the recognition of their desktop virtualization solution, I’m off to read up more, anything that can be done to improve the accessibility of desktop virtualization as a platform has to be a good thing.

Share and Enjoy

http://www.microsoft.com/technet/security/bulletin/ms09-sep.mspx

This bulletin summary lists security bulletins released for September 2009.

With the release of the bulletins for September 2009, this bulletin summary replaces the bulletin advance notification originally issued September 3, 2009. For more information about the bulletin advance notification service, see Microsoft Security Bulletin Advance Notification.

For information about how to receive automatic notifications whenever Microsoft security bulletins are issued, visit Microsoft Technical Security Notifications.

Microsoft is hosting a webcast to address customer questions on these bulletins on September 9, 2009, at 11:00 AM Pacific Time (US & Canada). Register now for the September security bulletin webcast. After this date, this webcast is available on-demand. For more information, see Microsoft Security Bulletin Summaries and Webcasts.

Microsoft also provides information to help customers prioritize monthly security updates with any non-security, high-priority updates that are being released on the same day as the monthly security updates. Please see the section, Other Information.

Microsoft have released their security patches for September, so as ever, check whether your systems are affected and organize which systems will need to have the patches applied. There are:

• 3 for Windows 2008 (critical)
• 2 for Windows 2003 (critical)
• 3 for Windows 2000 (critical)
• 3 for Windows XP (critical)
• 4 for Vista (critical)

One of the security vulnerabilities does depend on your version of Internet Explorer installed, do check it out.

Share and Enjoy

The IT directors’ replies were as follows:

• Time taken to get a new starter up and running

At least a week, the main challenges are around getting a desk and then having their account made, IT security seem to take forever, and then the pc might need rebuilt or might be below spec so a new pc needs to be ordered. One of the things I’ve been making moves towards is pre-building pcs and just saying when a new guy starts, give them a new pre-built pc, avoid this there’s Janet’s pc, which could be unhealthy, out of support or create debates.

One of the things the business are particularly bad at is notification, just last week I had a manager phone up asking where his new joiners’ pc was only to find out that they had logged a call that morning saying, new user starting today at 9am, please set up with no information in the call.

• How long does it take to rebuild a desktop?

Interesting you ask that question, it used to take about three or four days from the time the user logged the call, to the time the desktop was collected and then another few days for the desktop to be built.

With consultation from desktop support and helpdesk, pc rebuilds are now classed as an incident, an engineer will check the inventory and check the pc, then from the users’ desk kick of the rebuild using Altiris at lunch or when they leave at the end of the day, with a rebuild itself taking about 1 hour. All that’s left is the post build configuration, the user specific Outlook, desktop settings and profile work but again that can be done at the users’ desk.

I wanted to get away from the action of rebuilding a pc being part of the service request five year plan, there’s no rush kind of thing, the old service request involved too many tasks and extra bits which weren’t necessary:

1. Verify desktop health
   1. Collect pc from users desk and deploy spare pc and verify it works for the user
2. Deploy build (this had to be scheduled by the help desk manager)
3. Install extra packages and return pc to user
4. Configure pcs packages and specific settings

The problem being, one engineer might work on each task, creating duplication of effort, of communication “how far did you get with… what build did you load on…”. Now I expect one engineer to handle the issue on their own (unless there are issues), to effectively own the rebuild from start to finish.

• How long does it take to build a server?

It takes about five weeks, it’s a continual improvement project which I’ve got two teams working on.

1. The first is the server/network teams, I’ve asked them to identify issues with the operating system build, the build process and the work flow.
2. The second is to work with purchasing, with asset management and move them away from the olden days way of doing business. They buy a server like they buy a toaster. They would order three 1u servers of similar specification a week, each time requesting three separate quotes. I’m trying to agree that we now adopt a simpler process every Monday and Wednesday we request quotes for our two standard configurations, that way when someone requests a server, there’s no delay in getting a price, slowing down the tendering process.

Next we invited asset management into the world of CMDB, they now had access so, when a server is ordered and delivered, they allocate a name (click a new server button), they enter the asset number and the details, it’s a R610 etc, so that when the task to data center planning comes they know the model type and asset details to allocate space. At the same time though, when a new server name is created that emails the server team saying asset PETS0028975 a Dell R610 server name WSWP000687 has been created, so they know all the details ready for the build, they just need to check the call for the operating system and confirm with the manager.

It’s a never ending improvement project, the virtualization aspect is easier but we need to structure it more, it’s fine that we can deploy a server in minutes, but I need to control that from a budget, charge back and licensing issue, so we’re trying to agree a process on that.

Share and Enjoy

I got the chance to meet with an IT Director and took the opportunity to ask him a quick questions, three actually. They’re the ones I think underpin how your IT really performs.

You see one of the discussions I’ve been having is how important or useful those top level reports on performance/utilization actually are? 

I know we have to supply these kind of reports to answer specific business or technical issues, however I feel we should also be considering the real user experience, by taking three things that we can measure to illustrate how effectively we deliver. Anyway, I digress.

Here are three questions that I asked and why:

• How long does it take for a new starter to be configured and operational?

By that I mean the following: A desk, a working IT standard desktop/virtual machine, an account, access to their various applications and what is the average lead time quoted by helpdesk to a user, “sorry love it can take a few weeks for everyone to do their bit” – it’s unacceptable, particularly if we consider the employee/consultant cost, so you mean the internal onboard cost of a user might be £3,000 because your IT teams, your billing, your helpdesk process doesn’t work unless someone is constantly chasing calls?

On this basis it might be costing you a few thousand pounds or more per user joining in downtime.

• How long it takes to build or rebuild a desktop?

But before I hear the standard 45 minutes, ok 45 minutes, but does that include the packages, the layered applications, the security patches and everything else.?

If I take Jill a standard user from accounts, log a call and say “please rebuild my pc, it keeps crashing”, from the point of logging how long does it take for an engineer to visit/call and ask questions, debate it, collect it, obtain a spare pc, allocate and configure it, then take Jills one for it to be rebuilt and returned a day or two later. Might it be a week? A few days to complete?

Would business funds not be better spent replacing the hardware than rebuilding in terms of man day cost? Do I want to be paying £150 a day for a few engineers, a helpdesk guy and in user time when I can buy a Dell pc under an agreement for just over £400?

• How long does it take to build a new server?

The total day lead time in provisioning and delivering a server from the time when I log a call to help desk or send an email to the project team, “give me a DL380 G6 with 16GB RAM, 2 cpus and 4x300GB disks). As part of this, I’ll use the example, “I want to replace WSINT1982, our DL380G3 which runs one of our SharePoint sites”, how long would it take to obtain sign off, to get a quotation, delivery, space allocated, the operating system layered applications and application loaded? Weeks? Months? What are the technical barriers to delivery from asset to purchasing, from racking to middleware or operating system load, and what technologies and changes in process can be made to resolve this.

They might be small issues, ones which at a board/senior level don’t matter, but it’s the little things that underpin the big ones.

Share and Enjoy

Found these ILO documents which might be useful for reference and contain some interesting whitepapers, do check them out.

Quick jump to manuals by category…

• General reference
• White papers
• User guide

Share and Enjoy

HP.com

• Fixed an issue that would corrupt the iLO 2 firmware image if a Remote Console session was opened while the flash was in progress.  Additional information can be obtained from the following customer advosory:

 http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01850906

• Fixed an iLO 2 hang that would occur when the On-Board Administrator requested a server name change using a string larger than 32 characters.
• Fixed an issue where the iLO 2 would become unresponsive during a broadcast storm.  This could also cause the HP agents for Linux to seg-fault.

HP have announced version 1.79 firmware for their ILO2, check if it applies to your systems, and remember that having the latest drivers/firmware is the first thing the vendor (or service provider) will ask when logging a call.

Share and Enjoy

http://www.bladewatch.com/2009/09/04/how-old-is-my-server/

I’ve published the file which is in the previous post on the documents page, or you can download it here.

The concept is for those management/inventory reports where you’re trying to establish where you are with the server estate for consolidation/hardware or server virtualization project. They dates are of course a guide only, and referenced from the HP quick specs guide, they do not constitute the official manufacture date of your server, though HP or your service provider should be able to provide more information if required.

After publishing this Chris called me and asked what guide I used to determine if a server should be virtualized?

At the moment (though situations change as do standards), I’ve been thinking that anything older than a G4, shall we say 2004 should be virtualized.

At this point my business as usual mind feels horrified and notes that these servers work, they typically (if managed and maintained) need not represent a higher support cost than the newer servers and not everything should be virtualized – that 700GB file server might be best left as a DL380 G3 with 4x300GB drives rather than eating half the storage on my ESX solution. Anyway, many at this point would be horrified and state that everything can and they would be right in most cases, however, as a vehicle to moving towards that scenario, let me summarize my position:

• Servers of this age are fully depreciated or should be;
  • We have obtained value for money from the underlying asset – five years is a good run and virtualization could extend this at relatively low cost
• We need to examine the economics of running older servers;
  • Cost of upgrading components of servers older than 2004 will be relatively higher, as will support contract/warranty costs
  • Environmental efficiency issues – particularly if you take a x86 server from 1999 as a file server and establish a power usage to performance ratio 380w for 45GB storage for example.
• Specifications of the servers stated:
  • Covering as much of the server estate possible, without encountering those servers that might have a higher amount of RAM/Disk space etc to maintain the number of machines virtualized per ESX server. It’s fine to say we can virtualize a DL380 G5 with 600GB of storage, we absolutely can, but would I not rather get rid of 12 older machines that use more power, rack space and would consume less (as a percentage of disk space/memory and cpu) than a newer server?

What works for you will depend on your business, your internal standards and the individual applications involved.

Share and Enjoy