Help me learn to be the sysadmin I need to be!

June 13, 2008 9:34 AM

What do I need to know and how should I learn it? I am the "IT" guy at my company, a small supplier of building materials, with an office staff of about 10, and a warehouse crew of about the same. Problem is I'm underqualified.

Before I arrived here the network and all workstations were set up and supported by an outside consultant. When I arrived my primary responsibility was maintaining our web site, hosted elsewhere, and minor IT support when I could answer or fix immediate problems without having to call the consultant. I was A+ and Network+ certified around 4 years ago, but obviously thats not all that impressive, and really I haven't done much in depth there, haven't been building or fixing machines or doing significant network administration.

Unfortunately over the past year or so the consultant has gradually become less accessible to the point of being essentially useless, and I have recurring nightmares about something disastrous happening to our server or network and not being able to fix it, losing all of our business data, etc. I chatted with my bosses, the owners, last night and they've OK'ed me to upgrade my skills in order to make sure everything here is working well and secure.

Currently we have a single server running Windows 2000 server. The workstations are an amalgamation of Win XP pro and Win 2000 workstation. We have a 24 port switch tying it all together, and some networked printers, and a DSL connection to the internet. The server provides a shared directory, Active Directory services, DHCP, and hosts our business application server.

Given what we have my current plan is to self-study towards becoming an MCSA on Windows 2000 server. Some of the win 2k tests are discontinued, but since thats what we're running I figured I would work first on the tests which aren't discontinued yet, like 70-270 Admining Win XP, then working thru the self study books on Win 2k Server like 70-215 Adminning Win 2k server and 70-218 Managing a Win 2k Server network, as if I was going to try to take those exam. Then after I was up to speed on what we're currently running I might want to learn more on what we might want to move to in order to make sure our network is not too obsolete or insecure.

So, what I want to know from you technical mefites is, is my plan reasonable? Am I being foolish to consider studying for obsolete tests in order to fill out my knowledge of the stuff we already have? Is there a better way I should approach this?

Help! I feel overwhelmed and I live in constant fear that I'm going to come in one day and be told "nothing works, fix it" and not be able to.

posted by Reverend John to technology (8 comments total)

Speaking as an Oracle DBA, the one thing that any $ADMIN cannot get wrong is recovery. Test your backups frequently and make sure you're able to recover data. Downtime is one thing, but loss of data is another.

Document, document, document-understand WHAT everything is and what it does.

You might want to try building your network-participate in forums, newsgroups, etc. Read up on others' experiences. Learn from others, find people that you can communicate with in a pinch.

Read during your downtime, see if you can get another server to 'play' with.
posted by neilkod at 10:10 AM on June 13

The only problem with studying for obsolete tests is that, if that is your primary effort, you may not be well prepared to handle problems when actually faced with a bad situation. I say that, because, in most cases, a suddenly FUBAR server or network is usually the result of a failed hardware component, or a major malware infestation, and your way back to sanity is more a function of the reliability and depth of your backup strategy, than it is a matter of how well you understand Active Directory configuration.

So, I would suggest that your first step is to get into the current state of your server, and particularly of your backup and restore strategy, both for the server, and for your workstations. This might well involve getting some face time with the consultant who set up and has been maintaining your server, for some heart to heart about what their plans have been for a "bare metal" restore of your systems, in the event of a major disaster. You may find that there is no comprehensive plan for such contingencies, or that whatever plan they had depended on the availability of materials like installation media, data backup tapes, and hard copy configuration information which would have been lost altogether in case of a fire, or which may have just been contaminated to the point of being useless by a delayed trojan horse bit of malware.

At a minimum, you want to review your current backup strategy and products, to be sure that you are getting complete, restorable backups, and that you know how to restore from the backups, and that you actually do a test restore of some part of your systems frequently enough to be sure that your backups are good. You want to be sure that you are getting fairly current copies of your backup media to offsite locations regularly. You want to be sure that copies of any written configuration data (software license keys, administrative passwords, model and serial numbers of server and workstations including hardware options and configuration details, switch, workstations, etc.) you or someone else doing an emergency restore or rebuild of your systems might need, are also stored offsite.

You may also want to audit your anti-virus and firewall products, and look at how your backup archive strategy could protect you in case of unintentional data deletions or other operational problems. Many people, for example, do an annual, quarterly, or monthly "save and store" of their servers, before doing periodic disk cleanups and compressions. If your record retention policy is to have 1 year of order history, 3 years of customer history, and 7 years of accounting history, are you in an IT maintenance procedure that supports and ensures that, or are there 5 year old customer documents in your server directories? Don't get bogged down in fixing problems you discover at the outset, until you are certain you have a comprehensive backup and restore strategy in place, but do make note of any problems you discover as you find them, for future corrective action.

You may find, as many small businesses do, that you haven't any trustworthy backups, or any real ability to perform a "bare metal" restore of your systems in the case of a real disaster. You may find that the decrease in hard drive costs of recent years now mandates moving away from a tape based backup strategy to a disk based one. You may find that your backup software, or the strategy you've chosen for using it, sucks in a restore crisis. Getting to a point of comfort with your backups and restore strategies, and disseminating the information necessary to recover from such a bad situation to yourself, the business owners, and a knowledgeable third party consultant (in case you are unavailable at the time such a restore might be needed), should be job #1 for you. With that exercise complete, you'll sleep better, and have the time you need to improve the rest of your skills.
posted by paulsc at 10:50 AM on June 13 [2 favorites]

Assuming that if that day comes, you still have internet access, make sure you have the Microsoft Technet Forums in your browser bookmarks. If there's one thing 22 years in IT has taught me, it's that there are very few problems someone hasn't encountered (and hopefully solved) before you.
posted by JaredSeth at 11:04 AM on June 13

Okay this quite a question. Some advice:

1. Relax. No one expects a low-level support person to do the job of a consultant who bills your take-home daily wage per hour. You need to make sure you and your boss understand that you can't do everything and in fact will need some tutoring from this consultant. At the very least he should show you how the backups are done.

2. Backups. Do these. Take a copy of them home every friday or every 1st of the month. Your boss needs to understand that in the case of a fire everything is gone. Off-site storage of even old data is better than nothing. Verify your backups by doing test restores. Find out what you need to backup from Mr. Consultant. If Mr. Consultant keeps ignoring you then your boss needs to deal with this.

3. Continue to study the MSCA/MSCE books but also tinker at home. Make a virtual machine of the free/eval 2003 server. Get comfortable with installing this, updating it, creating a domain, etc. Grab another computer as a workstation and make it join the domain. Have different dhcp scopes. Install WSUS and configure it. Send out some random patch to your workstation. Dont waste all your time taking exams. Start doing the stuff now.

4. Learn to read logs. Windows does a lot of logging. You'll also need to read the logs of your backup solution. This must be checked daily.

5. Understand that you are using a 10 year-old OS for your server and if the hardware is that old then it might be time to upgrade to something new. You'll probably need Mr. Consultant to help you on this.

6. Google is your friend. If you google search (or search at support.microsoft.com) the exact phrasing of an error or something out of a log then you'll most likely find the solution for it pretty quick. Also, if the answer appears at experts-exchange.com just keep scrolling down until you see the answer. You dont need to subscribe to see those answers as the top half of the site suggests. www.tek-tips.com is good too.

7. Dont live in fear. If you cant do something make your boss know this. Dont let them saddle you with unrealistic expectations. Youre cheaper than Mr. Consultant for a reason.
posted by damn dirty ape at 11:51 AM on June 13 [1 favorite]

I've been away from systems administration for many years now so I'm not going to address the technical side of your questions.

I will however advise that it's very important for you to communicate with your boss and ensure that his expectations of you are reasonable. Don't be a hero, because you'll end up with no life, chained to the server. If you need help, ask for it. If you need training, ask for it. It looks like you're doing all these things now, but remember to keep it up because for some people it is easy to fall into hero mode where it's more expedient just to fulfill (rather than refuse) unreasonable demands.

The bright side of course is that this is a great opportunity for you to ramp up your skills. If you can document how much money you've saved the company by supplanting the outside consultant, you should feel very justified in asking for a raise / bonus at the end of the year.
posted by randomstriker at 12:04 PM on June 13 [1 favorite]

You learn by doing, there's no magic. The best single tip: Google is your friend. Google everything you don't understand until you do understand it.
posted by meta_eli at 2:16 PM on June 13

Backups are the most important thing. Get that sorted, and you can google, join forums, read TechNet, etc. the rest. If you have good, restorable backups and know how to use them, nothing is fatal.

If you can make a small home network to play around on, great. It sounds like you've got a mishmash of old and medium-old hardware and software, set up by someone else. That is the worst possible environment, but also the most fun.

I don't know where you live, but if there is some kind of Windows admin group where you can meet people face to face, network and pump them for good classes (don't bother with certs unless you want to change jobs and your current job pays all of it), help, advice and all, that can really help.
posted by QIbHom at 7:19 PM on June 13

First, get a new consultant. There are all kinds of people who will be happy to take your money and do good work. Your consultant is the expert, so he or she should ultimately be responsible for disaster recovery planning. Given your level of experience and expertise, you should only be responsible for carrying out the day-to-day disaster recovery aspects - changing backup tapes, checking backup jobs, making sure people save their data to the server and so on.

Damn dirty ape covered things about as well as you can. Monkey around with this stuff in virtual machines or on a spare PC. The best way to learn is by setting things up, breaking them and then fixing them.
posted by cnc at 7:31 PM on June 13