Updating Free Software in the Enterprise?
wallykeyster asks: "I'm an IT Director for a small private university in the U.S., and we are largely a Microsoft shop. We pay over $15,000 each year for our Campus Agreement so that we can upgrade the desktop OS to our version of choice, run Office, and have some Client Access Licenses. I would like to move to FOSS solutions, but I'm having trouble finding support for Enterprise management. For example, OpenOffice and Firefox (both of which I use personally) would be easy first steps, but IE is updated automatically via our SUS server (and settings pushed to clients via group policies) and Office updates will be included soon. How are other larger organizations (i.e. more than 200 desktops) dealing with software deployment and updates? Is anyone using Zen with Novell Desktop Linux?"
run all your applications off of an NFS server..
that way, you only upgrade one copy on the server.
-Dirtbag
......You're getting off EXTREMELY cheap. If you switch to a different OS, or OSS, you'll easily spend more than that (many times more) in hiring people to support the new infrastructure.
-Randy
I'm in the same boat where I work. I'm trying to get Firefox officially supported, the biggest sticking point is the lack of an easy method to push updates. I think this is one of the biggest reasons Firefox isn't widely deployed in the corporate environment yet, sure it's easy to install it yourself and update it yourself - but that's not a solution in a controlled environment.
Pay 45K per year to hire someone to manage a homegrown house of cards "solution" based on rsync, rpm, apt-get, crontabs and other such industry stalwarts.
I think the choice is clear!
Befor you all start shouting about a Debian repository and mounting /usr/local/from-server, please consider:
The question is most probably about updating free software on Windows desktops!
Do you have classes (in either IT, CS, MIS, or similar) that claim to teach real-world skills? If so, a project to automate such an effort would be a wonderful class project for you guys to undertake.
Even if the class isn't about IT, this project can be used as a case study - for example, a class about software methodologies and software lifecycle mangement - or even a business class evaluating build-vs-buy tradeffs.
I actually tried this *once*. It seemed like a great idea on paper. I would train people to do something useful, and they could go to potential employers and pretend to have some sort of useful experience.
It was awful, and I will never do it again. I ended up spending all my time fixing stupid mistakes, and it was more work than just doing it myself. Especially since the ultiumate solution was to convert to linux, and set up a server to dish out rpm's and schedule updates via crontabs.
Let's assume you have 300 PC's, that's $0.96 per desktop per week for IT infrastructure which appears to be working well. If you were having all sorts of problems I'd suggest otherwise but given the information you've supplied I'd argue that your dislike of Microsoft (or preference for FOSS) is getting in the way of your ability to think logically. Remember the first rule of engineering: it it ain't broke don't fix it
Not really, assuming: 1) You're paying students $8/hour 2) You work students 15 hours/week (they gotta study sometime) = $6240 I can do 2 students with room left over for a trained chimp (we'll assume $2,500 worth of bannanas and computer repair bill from feces thrown at computer). With 2 students you could probably upkeep a small university ok (say 150 computers per student) after that I'd put the students to work finding an automated solution.
...in bed
Unix upgrades have been centrally managed since some time before the dinosaurs, using tools like rsync and NFS shares.
Nowadays, with RPM and DEB package managers, you also have the option to put all packages on a central FTP server and then schedule an update using the native update utility eg. apt, rpm or urpmi.
So, my reaction to anyone claiming that there is 'no support' for Unix, or that Unix is 'hard to manage' or that Unix 'doesn't have enterprise tools'. Is one of incredulity - like where have you been the past 500 years, man??? Sleeping???
Oh well, what the hell...
Ummm, if that someone was at all knowledgeable in Linux, do you think that he would be asking Slashdot? Don't you think that he would already know the answer to this and be hacking happily?
He's a Windows shop manager and may or may not have Windows centric staff. He does not have an Linux people so, hiring Linux people for just the migration would probably cost him more than the $15,000. But, the cost doesn't stop there, he will need to have Linux knowledgeable support personnel for ongoing support. That means either train what he has now or hire other people. Either way, he is likely to run over the $15,000/yr. Microsoft tax.
I love Linux and use it exclusively, as I have for over four years. but, there are times when Linux is not the best solution and it sounds like this is one of them. $15,000/yr. is dirt cheap for Windows server licenses and CAL's, probably Exchange and SQL licenses, Windows Desktop licenses and Office licenses. It WILL cost him much more, at least for the short term(1-2 years) to switch. And, that all assumes that he truly can switch and that some application isn't going to force him to continue buying Microsoft anyway.
You're getting Windows and Office on 200+ desktops for only $15,000 and you consider that too much? Are you on CRACK? That's like quibbling over whether you'll pay three peanuts here or walk 1000 miles to pay two peanuts.
When I started my latest academic sysadmin job, we were talking about hiring a couple of students, as that's what they'd always done in the past. I kept putting it off because the systems were so screwed up that I had to spend a huge amount of time trying to figure out what kind of crack people had been smoking when they set them up and didn't have the time or energy to train anyone.
In the end, we bought a brand-new server (we needed the disk space, anyway) that I set up from scratch. I migrated the data and built a bunch of tools to keep everything up to date, then moved on to automate handling the workstations. I never did get around to hiring anyone else, and the only time I remotely regret that is when I'm taking vacation.
Hiring students to do things is a cool idea in theory. You get some extra help and some backup, and they get some experience that they can take to their jobs. Unfortunately, their lack of experience, professionalism, and often just general understanding of how things work mean that important things get broken or never get finished because they lack the knowledge or the time to do the work.
It gets really fun when you have a couple of people who think they know what they're doing making changes without telling anyone else.
These days anyone can set up a Linux box themselves if they want to learn about the OS and how all the pieces work, and I'm happy to help out where I can. But having students work on ``mission-critical infrastructure'', such as the departmental web server or faculty workstations, just isn't worth it.
Any MSI package can be deployed either to users or computers via Active Directory. There would be a bit of mucking about involved in creating new MSI packages for each update to software - but this is easier than manually patching a significant number of machines. I'm surprised that very few open source projects aiming to dominate the Windows desktop release their binaries in .msi format as this would simplify their distribution in AD environments.
Ok, but if you want people to use certain smtp server, force it at the network level, not app level. And why cant people change the homepage? If you have user accounts why not allow them to change it? If it's a public terminal you can make the config file read only to users. Windows does not help much to harden machines, Zen may work, but it's a work arround at best.
unfinished: (adj.)
Forcing the smtp server at the network level would work great if all your workstations are on a local network, but if you're managing users worldwide roaming around with laptops it's not really an option. Anyway, with the security policies for application/desktop configuration, you can choose whether you're setting defaults for a user that are locked (they can't change them) or unlocked (they can). A read only config file would also work but it's not that convenient to have to manage several dozen config files in /etc/skel and then have to script out changes to every user's account anytime you want to force a change on everyone. With Zen you can have as many different groups of users as you want and each group (or even machine if you want) can have a separate set of policies applied to it, minimum list of software packages to have installed, etc.... All policy is group based so if I have a group of office workers who only need OpenOffice and Firefox and shouldn't be able to make changes to their preferences I can put them in a separate group from my power users who I can trust not to set their incoming imap server to something non-existant and then complain to the helpdesk. You would never use Zenworks to harden a single machine, you would use is to manage at least several hundred out to several hundred thousand. You're absolutely right about it being overkill for just a few machines.