General Problems and Solutions

Image Delay

One of the unique aspects of our project was that we worked in constant close coordination with our supervisors, Art and Don, during our four weeks at Lincoln’s Challenge. Much of our work was contingent upon Art and Don first completing certain tasks, such as providing us with equipment. Another crucial task involved the creation of the image, which Art had begun working on prior to our arrival. However, due to problems with three programs, Access 21, Learning 100, and Meeting the Challenge, creating the image proved to be a very challenging and time consuming task. As Art explained, Access 21 needed server-specific information to install, Learning 100 required a product key to install, and Meeting the Challenge did not recognize the available disk space during the installation process. Our group was concerned that the software problems would prevent us from meeting our project deadline. After three weeks of unsuccessfully trying to resolve these problems, Art decided to create an image with the operating system, XP, but without these programs. His hope was that he could resolve the software problems shortly after our departure, at which point he and Don would install the programs on all sixty computers themselves. Because of this decision, we were able to complete our project by the first week of November, as requested.

Contradictory Instructions

Another problem stemmed from the fact that we often received contradictory sets of instructions from our two supervisors. Art and Don frequently took different approaches to various tasks. Their failure to consult with each other before giving us instructions often resulted in us having to redo an already completed task. For example, Don wanted us to use PVC piping to hold cables along the middle aisle. After we had already cut the PVC and duct-taped it to each desk, Art arrived and expressed his adamant opposition to this plan. He provided us with an alternative casing for the cables that was much easier to use. Another point of disagreement arose over wiring. Don wanted us to rewire the entire room to cut down on the amount of cabling we used because, as he explained, the longer the cable, the greater the signal degradation. Yet Art agreed with our assessment that we could use the wires that had already been fed through PVC pipes along both walls without risking a significant degradation in signal. While we appreciated the merits of Don’s argument, had we followed his wiring advice, we would have created significantly more work for ourselves than was necessary. Ultimately, because Art had the final say over what happened with the lab, we learned to consult with him before making any major decisions.

Group Communication

A third problem stemmed from communication issues between group members. With seven group members coming and going on different days, we realized that it was important to make sure everyone was aware of what was going on with the project. We decided to keep a notebook at the site that included daily entries detailing what had been accomplished that day, any problems that were encountered, and instructions for what the next group should do. Additionally, group members were in frequent phone contact with each other, particularly during weekends when we divided up into two groups; a member from the first group usually called a member from the second group to give them a verbal update. However, there were inevitable lapses in communication that led to some minor problems. For instance, when we were trimming the cables, we failed to properly delineate which cable along the windowed wall was going to connect to the hub; consequentially, this cable wound up being cut too short to make the connection. Fortunately, we were able to rearrange the cable connections and find another cable long enough to reach the hub. There was also some miscommunication over the labeling of cables. In one instance, a group failed to label the cables in the center aisle before securing them. This was easily resolved by using link lights to determine which cable came from which workstation. Additionally, one group used masking tape to label twenty cables after we ran out of labeling tape without realizing that another group member had purchased and created labels for those cables. This problem was also easily addressed, as the second group simply replaced the masking tape with the new labels.

Wiring Set Up

During our initial meeting, Rich and Art both expressed a desire to have the wiring as protected as possible to restrict students’ access to it. Given the setup of the room, this was a challenging task. We devised several strategies to accomplish this. First, we used cabling that had already been run through PVC tubing to wire all ten computers along the non-windowed wall, and eleven of the twenty-seven computers along the windowed wall. We ran the remaining sixteen windowed wall cables through a metal casing located a few inches above the floor. In the center aisle, we ran the cables through a protective casing provided by Art that we placed underneath the desks. We also trimmed the ends of all the cables to avoid having excess cable hanging off the end of each machine. In terms of the electrical wires, we did our best to make these as protected as possible. We used a covered extension cord to travel from the wall to the first workstation in the center aisle. We also used Velcro strips to bunch the cords together in this aisle and along the non-windowed wall so that nothing was dangling. Along the windowed wall, we placed the power strips up on a ledge to keep them off the floor, and we used a combination of Velcro and plastic ties to bunch the cords together.

Another wiring issue pertained to the huge bundle of hanging wires in the center of the room. These wires did not interfere with the network function in any way; they were just very aesthetically displeasing, and Art fervently longed for their removal. Fortunately, a maintenance person was able to come in and remove them (with Don’s assistance).

Cabling Issues

We encountered a reoccurring problem when making cables; several of the cables we made did not work, typically because of an improper wire order or a failure to adequately secure the plastic tip. Whenever we encountered a faulty cable, our solution was simply to make another one. We were usually able to discern whether a cable functioned properly or not by testing it with the two cable testers provided by Martin. However, on some occasions, we would discover that a cable deemed ok by the testers actually did not work when we plugged it into the computer and hub and no link lights appeared. Although we sometimes had to make new cables in these situations, we were often able to get the link lights to turn on by moving the cable to another connection within the hub. We’re not sure why this strategy worked; we did ask Art, but he was also unsure of why this happened.

One particularly frustrating cabling issue occurred at the very end of the project. We had tested all of the cables on multiple occasions and yet, on the second to last day, we discovered that the link lights were not on in Workstation 10. Because these had been on before, we assumed that the cable tip at the hub end had come loose at some point, as we had moved that connection around a lot. After we trimmed the edge of this cable and put on a new plastic tip, the link lights appeared again. We were very fortunate that this problem-solving strategy worked, for if it had not, we would have faced the daunting task of running a new cable through the PVC tubing to the non-windowed wall.

Specific Problems and Solutions

Workstations

One of our first steps was to go through and turn on all the computers, just to make sure that they all did in fact turn on. During this process, we identified problems with three workstations. Workstations 52 and 53 needed their hard drives to boot, and Workstation 31′s video card had malfunctioned. The first two issues were resolved by inserting a hard drive into the computer before booting. Because the video card in Workstation 31 was built into the motherboard, Don replaced the motherboard on this computer to fix this problem.

Electrical Issues

On more than one occasion when we were setting up the electrical wiring, we blew a fuse. We resolved this circuit-overload issue by spreading out the connections so that there were less cords drawing power from a single outlet.

Hub to Hub Connections

We were led to believe that there was a problem with the hubs after the link lights of the hub to hub cables failed to appear. However, after Art explained that these lights will only appear when data is being transferred between hubs, we were convinced that the connections were fine. Yet it turns out there actually was a problem with these cables, although Art did not realize this until after we had left. This problem is discussed in more detail below.

Cable Measurements

Determining the correct length for each cable was a crucial aspect of the cable making process. The measurement process was particularly tricky for the sixteen cables along the windowed wall, as they had to travel from the hub down to the metal casing near the floor, through this metal casing until they reached an opening nearest their ultimate destination, and then from this opening to the back of the computer. Although we took great care when making these measurements, we discovered that many were not accurate when we tried to run the cables we had created. We wound up remaking about half of these cables due to inaccurate measurements. Fortunately, we had extra cabling, so this was not a major set-back.

Server Cabinet

We ultimately decided to place the server cabinet at the front of the room because this is where the preexisting cords that went to the windowed and non-windowed walls originated from. So this placement made sense because we needed to connect these cords to a hub within this server cabinet. However, we discovered that the wires, which came out of a tube from the ceiling, were not long enough to reach the hub. We resolved this dilemma by placing the server cabinet on top of a wooden desk that was located at the front of the room.

Imaging

We encountered a myriad of problems when imaging. As this was the first time we did more than simply turn on the computers, during this process we discovered that there were some workstations with keyboards that did not work. This was easily resolved by fetching keyboards from the old computers and attaching those instead.

One of the floppy disks that contained the imaging software, Partition Commander, did not work. Art and Don tried to create a new floppy, but they were unsuccessful. This was not a major hindrance because we had two other working floppy disks, but it did slow down the imaging process somewhat. Fortunately, the floppy was only needed at the beginning of the imaging process, so the inconvenience was minor.

Another problem we encountered during imaging dealt with the boot order. In some instances, the computer booted directly into Windows instead of first reading the floppy disk. We resolved this quite easily by going into the BIOS and changing the boot order so that the computer booted through the A drive first.

During the imaging process, we discovered that Workstations 44, 49, and 59 had floppy drives that did not work. We determined this was the problem after we could not get Partition Commander to start on these computers, even after changing the boot order. This was problematic because our only copies of the imaging software were on floppy disks. Art and Don tried to create a CD with this software so that we could avoid using the A drive, but they were unsuccessful. Although we contemplated replacing these A drives, we did not do so at Art’s behest, as he said they are rarely, if ever, used by students. So we needed to find a different way to transfer the image to the hard drives of these computers. We ultimately wound up placing these hard drives in other machines with working A drives to do the image transfer. Once the transfer was completed, we returned these hard drives to their original machines.

After the imaging process was complete on a given computer, we had to reboot that computer to make sure the image had successfully been transferred. Two of the computers we rebooted showed a date and time error message, indicating that their CMOS batteries were dead. This message disappeared after we replaced these batteries with ones Art and Don fetched for us. Some of the other computers did not boot into Windows when we restarted them, indicating that the image transfer was unsuccessful. In these instances, we assumed that there was a problem with the hard drive (typically a Maxtor). We resolved this problem by replacing those hard drives with extra hard drives provided by Art and Don. In every instance except one (discussed below), this strategy was successful.

Workstation 42 proved particularly difficult to image. After our initial imaging attempt failed, we replaced the hard drive with one of the unused extra drives. When this did not fix the problem, we tried to image from a different master drive. This strategy also failed, causing us to believe there might be a problem with the motherboard (since the computer seemed unable to read any hard drives). Before replacing the motherboard, we decided to try imaging from yet another master drive to another extra hard drive. After repeating this process several times with several different masters and blank drives, we were finally able to get the image to work, although we are still not sure what the problem was with this computer.

Problems That Emerged After We Left

Hub to Hub Connections

It turns out there actually was a problem with the hub to hub connections. We had thought this was the case when the link lights for these cables didn’t appear, but Art reassured us that the lights should only appear when data was transmitting between hubs. However, we received a follow up email from Art two weeks after the project was completed in which he noted that some of the hubs were connected together with straight cable connectors instead of cross over configuration. In other words, the wire pattern at the tips of these cables was incorrect. Art took full responsibility for this error, as he realized he never told us that these cables needed to be made this way. He also said it was a very easy problem to fix; he and Don just needed to redo the tips on these cables to convert them into crossover ones.

Workstations

In his email, Art also said that Workstation 59 stopped functioning due to some hardware issues. This was unfortunate because this computer had been working properly when we left. Art did not say if he planned on replacing this computer with an older one from the previous lab or just getting rid of it entirely. He did note that one hardware failure out of sixty is pretty good, which is an observation I think our entire group can agree with.