latest news

contact info

Address:

Mathew S. McCarrell

Clarkson University

P.O. Box 3140

Potsdam, NY 13699-3140

Phone:

518-314-9214

EMail:

mccarrms@clarkson.edu

mccarrms@gmail.com

AIM:

CmikC114

Clarkson University Internet Teaching Lab

ITL Plaque

MP152 (1 Credit)

I am currently a member of the ITL's Lab/Server/Router Maintenance team and a member of the ITL 2006-2007 Executive Board.




Mid Semester Report (10/22/2006)

This semester I joined the ITL and enrolled in MP152 for one credit. My goals for this semester are to become actively involved in the Lab/Server/Router Maintenance team since I have little experience in programming and more experience in this area. I have currently been working mostly with Todd Deshane on most of the lab maintenance. I have taken control of the Windows XP Build and will be creating a new build within the next couple of weeks.

I am also in the process of decommissioning Cookiemonster (One of the Servers in the Lab) because it is ten years old, has a Pentium II Dual Core, the hard drives are held in each by only 1 screw (IDE Cables are too short), and it probably has very little ram . I am planning on using the server formulary known as Ernie (Previously a Windows 2003 Server) and the four 250GB (Possibly only three of the four because one may have been ruined by renovations in the labs this summer) to make a samba server for the labs. This will be a learning experience for me because it will be my first time setting up a Linux server.

One of my other projects I have been working on is Windows Server 2003. I have been experimenting with Active Directory to see if it would be useful to the lab and to allow us to learn to successfully ghost computers that are attached to a domain such as Clarkson.s AD. I am hoping that if there is a practical use for it then we may be able to use remaining money in the budget at the end of the year to buy a new Windows server.

I have been working with Todd on perfecting ghosting and network booting for the labs. One of the main goals in the lab this semester is to get network booting up and working so we can use it to ghost the lab. Currently we go to each computer with a CD in order to ghost them. Todd is hosting the net boot server as a Xen Guest on Snuffleupagus. We currently use UDPCast to ghost the labs which is more efficient than G4L and allows us to push out changes in the current build more often. We are also in the process of writing both a Linux and Windows script to reboot all the computer in the lab so that they can reboot and receive the UDPCast file from Snuffleupagus with out us having to manually restart them all. We are also planning on using G4U to make a master backup image when we do a new build for the lab. This will allow us to keep an image that is not modified by other users so that we can just update the image and ghost it out.

So far there have been lots of progress in these areas. The network booting server should be correctly working in the next few weeks. There will be a new lab build for Windows (possibly Linux as well) in the next week or two. The networking in the lab (which is currently awful) should be redone before the end of the semester which will also make it easier for us to network boot because we can run our own network on the second network cards in all the computers in the lab. I may also be participating in a team that will be redoing the networking in the labs if the labs decide that it is too expensive to have it professionally done. My future goals are to further increase my knowledge of Windows which will be expanded with the addition of Windows Vista and Windows Server and to learn as much as I can about Linux.

End of Semester Report (12/13/2006)

This semester I learned numerous things about maintaining a computer lab such as the ITL. I ended up scraping the new lab build that was planned in early November because Thanksgiving break came up very fast and I had no time between homework and exams to finish it. I am currently planning on coming up early before the beginning of next semester to do a new lab build.

I am planning on implementing many things on the next lab build for next semester. I am hoping to be able to implement the Windows XP Shared Computer Toolkit which is designed specifically for use in public labs. The toolkit basically allows you to keep the build on the Windows partition preserved and grants the user only a previously allocated separate partition to work on and is erased upon restart. I think this solution will be easier to implement than a DeepFreeze Trial I was thinking of implementing; however, I may still try DeepFreeze out on a limited number of computers so see which computers appear to run smoother in the lab. Pat Wilbur and I recently discovered about two weeks ago that we can ghost over COSI's build that since the computers between the labs are so similar it works. Therefore I am planning to include COSI's Fedora Core 6 build on the lab machines next semester as well. Mostly I plan to do this because I can and because it is linked to AFS which many people prefer to log on too rather than using ssh into polaris.clarkson.edu. Todd Deshane will also be setting up the Ubuntu Linux build while I watch to learn the process.

I have been working constantly at perfecting the ghosting of the lab. I will plan to experiment with sysprep, a tool that allows a computer to remain on a domain after ghosting it, next semester. My main reason behind waiting is the ITL's lack of its own internal network that is separate from Clarkson's network. Once we have a separate internal network, I will be able to set up and Active Directory for the computers. The reason I have not set it up on Clarkson's network is due to problems I encounter with Clarkson's DNS server. I am also hoping that Samba 4 will be released next semester so I can test how they are emulating Active Directory. Samba 3 only works with Active Directory for permissions and file sharing currently. In searching the UDPCast site, I also discovered versions of UDPCast that run on Windows and Linux while the machine is running. I have found this very usefully because it allows me to make a backup image of the build at the times when we ghost the whole lab. It is also much faster than using G4L for this method since it does not use a FTP server. I am also currently working on having the UDPCast kernel be an option in the GRUB boot menu that appears upon start-up on the lab computers. This will allow us to ghost without having to run to each computer and insert a CD and will provide a backup method of ghosting in case our future net boot server goes down and we are unable to net boot UDPCast.

One other project that is ongoing is setting up my first Linux server in the lab which will be a Samba server. Previously, I had planned to have one up in about a weeks time; however, I ran into some complications. Upon reassembling the server after cleaning it, I was unable to get it working correctly with the new hard drives. I later discovered that the IDE cables were weak and that was causing most of my problems. There may also be one hard drive that is failing and giving me problems as well. I had to put the project on hold because of how busy the second half of the semester got for me. Most likely it will only take a short period of time to get it up and working correctly. I am currently planning on possibly doing this before I leave for the semester, otherwise I will do it first thing next semester.

There are many other numerous projects that I will be involved with next semester. I am planning to be part of the team that Pat Wilbur wants to make to manage the storage policy and backup policy of the labs. This will include allocating the 8.5 Terabytes of RAID storage that COSI received from IBM with the Blade Center. I am also planning on remaining on the lab/server/router management team. Hopefully by next semester the labs will have the new networking installed and I will help manage the internal network. Our internal network will consist of two 48 Port Linksys Gigabit Switches and our own DHCP and DNS server that we will set up. We will also be expanding on the Clarkson network in the labs by adding two 48 port 10/100 mbps switches to help run the approximately 100 Ethernet jacks that will be in the lab. I will most likely try and be involved in the management of the Blade Center once we have power to run it. Hopefully, I will learn a lot about Linux and Xen through being involved in this.

This semester I feel I have accomplished many good things. I am heavily relied on by Todd to maintain the lab. I have also gained knowledge on maintaining different parts of Ubuntu Linux distribution as well as learned a few new things about Windows that I did not know before. I am greatly looking forward to many of my upcoming projects for next semester; however, many of the projects I will be involved in are relying heavily on the new networking being run over winter break. My major goals for my future in the ITL are still to learn as much as possible about Linux, further increase my knowledge of Windows, learn more about virtualization, possibly give a presentation about advanced features of Windows, learn HTML so I don.t have to make this page in Microsoft Publisher (This page has sence been updated), and continue to improve the labs in general.

Windows Scripts (Converted to .txt to allow view in multiple browsers)

Walkthroughs

Important Links

Back to McCarrMS Homepage