The layered picture of a computer in 0003.b has a practical payoff when something stops working, because a problem lives somewhere in that stack: in a cable or a part, in a driver or an application, or out in the network. Troubleshooting is the work of finding where. This section first describes a general process that applies to any problem, then follows it through the kinds of fault a teacher is most likely to meet, roughly from the physical layer upward: power and connections, faulty components, software after an update, and network connectivity. It then turns from fixing to preventing, and ends with a different kind of boundary, the line between what a teacher can fix and what belongs to an organization's IT staff.

A general process#

Effective troubleshooting follows a sequence that applies to any problem.¹ First, identify the problem: confirm the symptom, so you know exactly what is and isn't happening. Then ask what changed recently. A new problem often follows a recent change, such as a moved computer, a new cable, or an update, so this question narrows the search, although the suspected cause still has to be tested. Next, start with the simplest and most likely causes. These are common and quick to rule out, so a list of remedies that begins with replacing hardware and ends with checking a cable should be reversed, trying the quick and common remedies before the costly ones. Change one thing at a time and test the result after each change, because if you change several things at once and the problem disappears, you can't tell which change fixed it. Finally, document the solution. A record doesn't stop the problem from happening again, but it lets others, or you, resolve the same problem more quickly next time.

Power, connections, and settings#

The simplest and most likely causes of a hardware problem are usually the most physical ones. Before assuming that a part has failed, check power, cables, and settings. A projector showing nothing may be set to the wrong input source, or its cable may be loose. Speakers that make no sound may be fine, with the volume turned down or the wrong audio output device selected. Equipment that stops working after being moved has often had a power cable pulled loose or its power strip switched off. These checks take moments, and they can save an unnecessary purchase or service call.

Isolating a faulty component#

When the simple checks don't solve a problem, the next question is whether the fault lies with a component or with the system it is connected to. The way to find out is to isolate the fault: swap a part, or test it on another system, and see whether the problem follows the part or stays where it was. A monitor that shows nothing can be connected to another computer known to work. If it works there, the first computer is at fault; if it still shows nothing, the monitor is. The reasoning runs the same way in other cases. If two different headsets fail on one student's laptop, and one of them works on another computer, the evidence points to the laptop. If a memory module that causes errors in one computer causes the same errors when moved to another, the fault followed the part, and the module is likely faulty. Some cases need two tests: to tell a faulty keyboard from a faulty port, try the keyboard in another port and another keyboard in the original port. As in the general process, changing several parts at once hides which one was faulty.

Software after an update#

Software problems often follow an update to the operating system, and the layers in 0003.b suggest where to look. If a device such as a camera stops working after a system update, the update may have left it needing a compatible driver, so a good step is to install an updated driver from the operating system or the manufacturer. If only one application fails, the failure is confined to that application, which may need an update of its own to work with the new system, so check for a patch or reinstall it. If the manufacturer has already identified the problem and released a fix, install it. When no fix is available yet, rolling back the update can restore function while the problem is reported and resolved. This risk is also why IT staff often test updates on a small group of computers first, so that an incompatibility is found before it affects everyone in a school.

Network connectivity#

Connectivity problems call for the same narrowing, first by device and then by destination. Check whether other devices are affected: if they are, the problem is in the network, and if not, it is in the one device. Restarting the device and the router (and the modem that connects to the provider) clears many temporary faults, and confirming the device's network settings rules out a misconfiguration. Then test whether the problem involves one site or the whole connection, by trying other websites on the same laptop and that site on another device.

A few specific symptoms point to specific causes, using protocols described in 0003.e. If sites load when you enter their IP address but not their name, the connection is working, but names aren't being turned into addresses, so suspect the Domain Name System (DNS). If a device shows an address beginning 169.254, it has assigned that address to itself, which suggests that it got no address from the DHCP server that normally hands addresses out. If everything on the Internet fails at once but computers can still share files and printers on the local network, the problem likely lies in the connection to the outside, so restart the router and modem and then check for an outage at the provider.

Prevention#

Many problems can be prevented, and the habits that prevent them are routine. Regular updates close known security flaws, so delaying updates for months leaves those flaws open. Backups to a separate location let data be restored after a failure, which no other measure does. Systems need adequate free storage, since updates and temporary files take space. Surge protectors between computers and wall outlets reduce the damage that power spikes do to equipment. Keeping devices clean and cool, for instance by cleaning dust from computers and improving a room's ventilation, prevents overheating.

Working within restrictions and knowing when to escalate#

The steps so far assume you can change what's broken, but in a school many settings are controlled centrally, and a teacher often lacks administrator rights, the permission needed to install software or change system settings. These restrictions exist for security and management, so the right response is to work within them. Approved browser-based tools can meet short-term needs, because they run in the browser that is already on the laptop and nothing new must be installed. A teacher who can't install a file converter can use an approved online converter instead. Such tools should be tested before a lesson, so the lesson stays on track. They also send data to outside servers, so check their privacy terms before uploading any student data. Using a workaround to get around a restriction, by contrast, circumvents a security control and may violate school policy.

Some problems are beyond a user's control altogether. Network restrictions such as blocked sites are managed by an organization's IT staff, and so are building-wide outages. These should be escalated, which means reported to IT staff, with clear details that let them reproduce and diagnose the problem: location, scope, time, the exact error message, and the steps that produce it. "Room 204 laptops cannot reach example.org since 9 a.m.; other sites load" gives staff far more to work with than "the Internet is down." If a site needed for a lesson is blocked, ask IT staff to review the block, giving the site's address and the lesson's need. Any screenshot sent along should be checked so that it doesn't reveal student information, since student data is protected and what is shared should be limited to what is needed. Several of the network symptoms in this section only make sense once you know how a network is built, which is the subject of the next section.

• isolating a fault: swapping or testing a part to learn whether the problem follows the part or stays with the system
• rollback: returning to the version before an update to restore function
• backup: a copy of data in a separate location, used to restore it after a failure
• surge protector: a device that shields equipment from power spikes
• administrator rights: the permission needed to install software and change system settings
• escalation: reporting a problem beyond a user's control to IT staff, with clear details

Practice: practice questions for 0003.c

1. MWC1, Typeface: decomposition and debugging in a problem set. ↩