I have noticed a nice spike in traffic on all three of my blogs (Kevin Mullins, The Tech Manager, The SysAdmin) an wanted to thank folks for reading and remind everyone that comments are welcome. In particular, my SysAdmin blog has really come on strong with twice as many hits and page view as my other two blogs. Again I thank you and remind folks that I am always open to share ideas and strategies on all three blogs and welcome your help and participation.
For those of you who have not visited in a while, I moved my "Technical Manager's Perspective" blog to http://www.techmgr.net and I have started another blog at http://www.syadmin.net
I am coming up on my one year anniversary of blogging and I am looking to make a couple changes in focus and strategy with my blogging.
I really enjoy blogging and communicating and I know that I have a little following, however most of the folks that have communicated with me are more technical non-management folks and my interest and/or search for improvement is more on the managerial side, which leads to more posts about managerial content as opposed to technical content.
So, in an attempt to maintain both a technical and managerial focus I have created an new blog for system administrators called The System Admin which is located at www.syadmin.net. Notice the spelling and play on sy.
I have moved my Technical Manager's Perspective blog from my www.kmmm.net domain to www.techmgr.net. I am not sure what I am going to do with my www.kmmm.net domain however I will be focusing on both the www.techmgr.net and www.syadmin.net domains in the next few months.
I have responsibility for both the Unix Operations and the Windows Operation within our company, and I struggle to find tools that work well in both environments. Both groups work autonomously and do a great job of supporting their respective environments, however as our responsibilities grow, I would like to find consistent tools for maintaining, monitoring and supporting our environments.
Two tools that do work in both environments are BMC Patrol for performance gathering and HP Openview for event notification and monitoring. We are not using Openview to monitor our network, instead we use it for system and event monitoring. We have a decent baseline of monitored events, however we struggle with event escalation especially when new errors or a new event occurs. We have installed IBM Director on the IBM nodes and it passes Hardware events to Openview, and we have scripted certain events which also hand off to Openview. There still is a lot of refinement needed in our process as many events enter Openview and are not reacted to because the criticality level is identified incorrectly or the followup and escalation are not defined in Openview. We are looking to reduce the number of system outages that impact our user community by improving our event escalation and notification.
Another tool that is in use in on our Unix platform is Tripwire.Tripwire is an auditing tool that will identify changes in files and will enable you to better track changes in your environment. We have started looking the Tripwire on the Wintel side and hope to have it deployed within a month.
I am open to others thoughts and experiences surrounding event monitoring and event notification in a multi-OS environment and would be glad to discuss this in detail.
Most IT managers are aware of the change in Daylight Savings Time and the impact this will have on our environments. I think we can all agree that it is not going to crash our systems or impact us in a manner similar to that predicted for Y2K, however our challenge as Technical Managers is to ensure that our computer systems remain up, available and in sync from a time perspective.
These days with enterprise wide domain controllers, email, calendaring, blackberries, and multiple gadgets you do not want to have a problem with time synchronization.
Here is a list of some of the critical patches you may need for your environment:
Good Luck .....
Last month I wrote about Microsoft's Virtual Lab Express which is a great tool for previewing and testing software.
As I have investigated further, I think I prefer the Microsoft Virtual Hard Disks (VHD), which are very similar, however the advantage is that you can download the VHD's to your environment, share them with your end users and then test the functionality in a very secure environment.
I feel that this is the new paradigm for software testing. You can create and entire virtual network within a VHD, along with the Active Directory, DNS, Front-End Application environment and Back End database server environment. You can use this environment for development, user acceptance testing or even offer this environment as a demo or testing environment to your customers.
If you are an IT professional with Microsoft based responsibilities, you should check out both the Microsoft's Virtual Lab Express and Microsoft Virtual Hard Disks.
Working in Healthcare, you quickly realize that there are few adoptions of bleeding edge or newer technologies, instead we are lucky to have and deploy three year old technologies and strategies. We recently deployed a virtualization strategy that was piggy-backed on a Blade purchase. The Blade purchase was approved in 2005 for 2006 and we were allocated limited funds for a Blade chassis that would be populated with 6 or 7 servers.
After researching our options in early 2006, we realized that many folks were using Virtualization on Blade technology and we started looking at Virtualization solutions. We compared both Microsoft Virtual Server and VMWare ESX server and found the VMWare product to be the better product. We quickly realized the VMWare's ESX server was very stable and deployable, and we proceeded to spend funds allocated for the Blade project on virtualization.
We quickly realized that it was more cost efficient to deploy the VMWare 2 CPU licenses on a robust standalone server with alot of memory, than on a Blade server with limited memory and disk capacity. You can deploy more Virtual servers on a beefy server with 16 or 32 gigs of ram than you can on a Blade server with 8 to 16 gigs of ram. We proved the merits of VMWare from a cost perspective and a deployment perspective and expanded our resources in that area.
This deployment could not have happened if we did not have the Blade funds, however one thing that I have learned is that sometimes you need to be creative to get your job done. We spent more on Virtualization by spending less on the Blade infrastructure and in the long run this benefited the company.
At Oracle World last month, Chief Executive Larry Ellison told thousands of attendees at the Oracle OpenWorld conference last week, "If you are a Red Hat support customer, you can very easily switch from Red Hat support to Oracle support." as reported by CNET. This story has generated a lot of controversy in the Linux camps of the world, and is one of the top story's on digg and Linux Watch.
This direction illustrates the power that Oracle has in the Marketplace. Oracle is doing this to ensure another stable platform for its products, however it comes at the cost of RedHat. How would you feel if you were a developer at RedHat or a contributer to the RedHat distribution. I think one strategy that Oracle could take would be to add resources to both the RedHat and Oracle-Customized RedHat distributions. Oracle could and probably will reduce their Licensing structure for Linux based products in an effort to drive adoption to Linux.
Someone who is in favor of this new Oracle strategy is Dana Gardner at ZDNET who wrote ...
"If Oracle can subsidize the offering of subscription services to an acceptable and open (not too forked) Linux distribution (and these need to happen) through its commercial products' revenues, it will. It was okay for open source service providers to undercut Unix and Windows via their low-cost development and distribution means, right? What's good for the goose …" Follow this link for the complete article
Will current customers really move to Linux ?
For larger Enterprises with Oracle databases, this path needs to be cost effective and supportable. Oracle is trying to bring some credibility to the support side of its products on RedHat, and this may be an good strategy, if they really can provide the support. From a cost perspective, it is almost an even comparison of Hardware and OS choices. You still need to purchase RedHat and a support contract for your products. Hardware cost of Database Platform Intel based servers and Sun Servers are very similar.
So, in my opinion, there is not a compelling reason for Large Enterprise Organizations to move your production database structures to Linux, however this could be a nice strategy for Small and Medium sized businesses.
We have an older Pioneer DVD JukeBox that stores over 200 DVD's. This device was installed over 5 years ago and all of the folks involved in the installation have left our group. It was fairly dormant for the first 2 or 3 years however about 2 years ago we started using it as an archive repository. We use Point Software to virtually bind a number of DVD's together which then appear as a Share for folks and applications to use.
Over the last 6 months we have had a number of hardware related issues with the Jukebox. Internal switches would fail causing problems, shares would fail to mount and most recently two DVD drives had to be replaced. With this most recent failure we also realized that the bad DVD drive, damaged one of the DVD's in the share. There were 40 DVD's in the share and it would not mount because one of the DVD's was damaged. You could actually see the damage on one side of the DVD so we pulled the damaged DVD and worked through remounting the share with only 39 DVD's.
We tried to read the damaged DVD on multiple DVD readers however we could not read the DVD. We realized that we did not have a backup copy of the DVD, breaking Kevin's Rule Number 1 for all System Administrators, Backups are a Requirement. After further investigating, the original thought was that we backed up the data before archiving, so we would not need to backup or copy the archived data on the DVD server. However, with data spread across 40 drives, it was hard to determine which data was missing and a further complication was the fact that most of the data was from 2003 and we had selective backups from 2003 that probably would not work.
We were in trouble and Directors and VP's were looking for answers. I Goggled "Data Recovery" and identified thousands of companies that claimed they could restore your data. I called 4 different vendors and chose one close to work. I had a pay a premium to escalate our recovery however the vendor turned the restoration around in 2 days and recovered all of the data on the DVD. This allowed us to move the data back to the archive share however stressed the need for a backup or copy of our Archived DVD data.
Once the crisis died down, my team investigated how to copy a DVD within the Jukebox, and has started the process to make a duplicate DVD copy for all Archived DVD's in the Jukebox.
Though-out this process, all I could think about was RULE Number 1:
You may have read my post from June on System and Network Administrators Tasks along with my little update to that post in July. If you happen to miss that post, I identified the tasks that I feel every System and Network Administrator need to understand to become better administrators. I am republishing this list, as I would like to touch base on a couple of topics related to this list over my next few posts.
1. Backups are a requirement and are crucial to your success
2. Monitoring Tools
3. Maintain Service Contracts with all vendors
4. Maintain Version Control
5. Consistent Hardware Builds and Software Installations
6. Documentation
7. Communication
8. Teamwork
9. Training and Schooling
10. Stay Current with Technology
11 Know your users or your customers
12 Know your line of business
