Slashdot Mirror
Can Windows, OS X and Fedora All Work Together?
greymond writes "In my ever growing job responsibilities, I've recently been tasked with documenting our organization's IT infrastructure, primarily focusing on cost analysis of our hardware leases and software purchases. This is something that has never been done in our organization before and while it's moving along slowly, I'm already seeing some places where we could make improvements. Once completed, I see this as an opportunity to bring up the topic of migrating the majority of our office from Windows 7 to Linux and from Exchange to Gmail. However, this would result in three departments each running a different system: Windows, OS X, and most likely Fedora. Has anyone worked in or tried to set up an environment like this? What roadblocks did you run into? Is this really feasible or should I just continue to focus on the cutbacks that don't require OS changes? (The requirement for having three different systems is that the vast majority of our administration, who rely solely on an install of Microsoft Windows, Word and Excel, are savvy enough that if they came in and saw Gnome running on Fedora with Open Office they'd pick it up fast. However, our marketing department is composed entirely of Apple systems, and the latest Adobe Creative Suite doesn't seem to all work under Wine. The biggest issue is with the Sales department though, as they rely on a proprietary sales platform that is Windows only — and generally, sales personal give the biggest push back when it comes to organizational changes.)"
Why do you want to get rid of Exchange for GMail? What has it not been doing for you? I'm at a small company, and we have Macs, Windows and Fedora desktops. The only changes we've made was removing Office for Mac and replacing it with Mail.app on the Macs and using OpenOffice on the Macs and Linux desktops.
All tied together with the an Active Directory on Server 2003 and an Exchange server.
"I use a Mac because I'm just better than you are."
[1] Glossary:
Boni: plural of Bonus.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
i recently hired an IT staff that outsources their job responsibilities to online chat message boards. has anyone else had experience in replacing such a staff?
No.
What is Windows 7 failing to do for you that Linux will improve upon without causing problems in different areas? I find it hard to believe that a business that already paid for Windows 7 is making a smart business decision by dropping it in favor of Linux (or even Mac OS X).
Changing to Linux because you can is just stupid. Good luck following through with your "savvy" users actually using Linux on a daily basis without a lot of trouble. You're going to need it...
Post pictures of your girlfriend, and we'll tell you if you should propose. Give a snapshot of your kitchen, and we'll make redecorating suggestions. Post your eTrade login and password, I'll take a shot at helping you revise your portfolio. Thinking of buying a house?
We know nothing about your company, what it does, what the people are like. We have no fucking clue what you should do, because every situation is different. If there is one decent bit of advice to be had, and this comes from the Veep level with 20 years in:
1. Everything starts with the directory system and
2. Calendaring derives from it.
Long answer: Noooooooooooooooooooooooooooooooo!
XML is a known as a key material required to create SMD: Software of Mass Destruction
When was the last time Gmail was taken down by a virus? Or a power outage? Or a hardware failure?
I Am My Own Worst Enemy
Lets break this down: "I see this as an opportunity to bring up the topic of migrating the majority of our office from Windows 7 to Linux and from Exchange to Gmail" -Why? Most users are not comfortable with anything other than Windows. Second Windows 7 is still somewhat fresh, I mean, your going to depreciate software that you likely purchased less than 6 months to a year old? Sounds like an immediate waste of money rather than a long term savings. The Second part of your question makes it seem like your some dumbass fresh out of college. Really? GMail over Exchange? Are you willing to hedge your business needs on a free email service, not to mention the loss of collaboration options, etc? All in all it sounds like this situation: 1) You're a Junior Administrator or Helpdesk Engineer 2) They fired the ACTUAL IT Staff and left you since you are cheap enough to keep on the books. Look, want to save money? Look into Virtualization Options, Open Office instead of Microsoft. Linux is not the end all at the workstation level, no matter what they tell you in college.
When was the last time Gmail was taken down by a virus? Or a power outage? Or a hardware failure?
http://www.pcworld.com/article/160153/gmail_outage_marks_sixth_downtime_in_eight_months.html
This all depends on the size of your network and number of each type of system deployed. Plus don't forget there are political reasons for making or not making certain recommendations that generally outweigh any technical/economic reasons. I have seen people fired for making recommendations that had less exposure than what you have suggested.
Encryption: I may not agree with what you say, but I will defend your right to encrypt it...
Virus, power outage, or hardware failure? Not sure. Unexpected outages? Well, at the very least, 2009. I'm sure there have been at least local outages in 2010, too.
Now the fun part starts: how much would it cost your company to make your mail service as reliable as Gmail? And from the fine article posted by the AC above:
It may sound bad, but Gmail does appear to have a reasonable amount of uptime, all considered. Following last fall's series of outages, a Google rep told the IDG News Service that Gmail suffers only about 10 to 15 minutes of downtime per month, giving it an average uptime rate of 99.9 percent. He noted that, according to some independent reports, on-premise e-mail systems tend to see twice the amount of offline time--anywhere from 30 to 60 minutes, on average, every 30 days.
Is Gmail for everyone? No, but it certainly is worth looking at for some companies.
I Am My Own Worst Enemy
Management do not want big changes. They want quick wins. Find somewhere that can show savings fast. If you find several, keep some for next years savings. And sometimes management lose attention to the issue, so talking is enough. Then you can use the same savings next year. Especially if management change. Hell, we presented decommissioning the same server 3 times to various management. Happy managers all the way!
And whatever you du: Do NOT propose anything that require more work. You will not get more staff. You will not get more time to do it. In the end you will be the one paying for the savings.
Sorry mate, but some of the advice you give is rubbish:
- "more professional to have a @companyname e-mail over @gmail."
You do know you can use google apps for your own domain, right?
- "I don't know if you are currently using or plan to use active directory"
You do know that Active Directory is a requirement for Exchange, right?
http://www.zombieapocalypse.tv/
The biggest issues are probably the file servers (NFS is only allowed for the default Ubuntu install, Samba for everything else)
Is that right? I am pretty sure that Ubuntu Desktop can view Windows file shares with the default install. Or do you mean on the server end? Yes, you might need to install Samba in order to have Ubuntu file servers support Windows clients, but it's not particularly hard.
The bigger and more annoying problem that I've had with file servers supporting different client operating systems has been that the different systems treat metadata differently. Different operating systems have different methods of dealing with file permissions. Moving a file might not keep your old timestamp. Windows puts Desktop.ini and Thumbs.db files all over the place, and OSX puts .DS_store files and resource forks everywhere. Moving OSX files from a non-OSX system can still cause you to lose resource forks, which isn't generally a huge problem, but it's annoying.
printing (maintaining both Windows and Unix print queues is apparently difficult).
Again, my recollection was that I was able to set up Ubuntu desktop to use Windows print queues. Maybe I'm forgetting something.
Why Linux? If it's simply license costs, well then keep people on Windows. The per-seat software license costs are pretty small compared to your labor + overhead costs of what your IT people will need to put in to retrain user expectations. Even if you're paying $500/user for Windows + Office, that's tiny compared to overall productivity differences.
If people need posixy goodness, give 'em OSX. For the most part they'll probably be happier to not need to mess around as much with desktop config and software installation. Leave Linux to users who can self-install and self-support.
Do not take MS Office away from your Finance and Management teams. Sure, they could learn OpenOffice if they needed, but there's a lot of stuff that Excel does really well that OpenOffice Charts can't. And if a Senior Manager spends even 1-2 hours trying to learn how to use OpenOffice, well, that wasted time just blew away the license cost savings. Re-training and loss of productivity is very expensive, very difficult to factor into your budgeting plans, and impossible not to underestimate.
Finally, why move from Exchange to GMail??? If you don't want to pay as much, consider Kerio or Zimbra, but do not force users to give up integrated messaging, group calendars, and contact databases. We're moving right now from a lousy group calendar to Kerio (Exchange wasn't right for us) because we waste so much time just trying to schedule meetings.
I'm out of my mind right now, but feel free to leave a message.....
Difference for the sake of difference is not progress. Unless you're improving something, don't force your users to waste time learning a new system. If you've already paid for software that people are getting use out of, just leave it alone. This is one thing that frustrates me with a lot of technology companies, they just innovate in circles, recreating existing features and rebranding the same old services, merely making things different and forcing their users to adapt to a new system that offers no significant benefit.
Employee productivity should be a major goal of any good corporate IT force. Not all problems have technological solutions, many have human solutions. You need to include the human factor in your problem solving, and if this means sending out an e-mail asking for feedback or walking around the office talking to folks about what problems they encounter and what features they don't understand, then do it.
This is a main difference between an IT department that people hate, and an IT department that people love.
The last transition I ran (had to leave due to personal reasons) was looking like it was ultimately going to fail.
Why?
OpenOffice - found several critical bugs (all fixed now) that kept people from being able to work effectively
Intel video drivers - found a fun critical bug whenever they plugged into a projector
Didn't have control over what other groups bought as software (big one, make sure management is actually willing to back you up)
* think hard about this one, is there anyone (manager) in the company that will end up buying something without consulting you and who no one wants to go against...
The 3 OSes can easily coexist. Here's how I would go forward:
Don't touch the different platforms at first, start with the applications.
* Web browsers - make sure everyone is running firefox. I found out that 1 person was using IE6 for an important project. they hadn't mentioned it, even when asked directly. Solution: Block Internet explorer access, (I forced the person to move to IE8, yay for small victories)
having people complain when you have it blocked on Windows is much better than having people complain when they are now on Linux. (They will blame Linux)
* Best in class applications - DON'T start with OpenOffice. Make open source applications a regular part of discussions for new software. Evaluate other software you use for open source applications. Make sure they are successful.
* Make sure the other people in IT actually want this change.
* Move them to Linux/OpenOffice and observe problems over at least 1 full release of Fedora, trying to get problems fixed for the next one
* Transition office to OpenOffice on all machines (have just installed first, then default, then uninstall MS Office - very important) watch for issues over at least 6 months
* Transition office to Linux
Yes, this is more like a 2 year plan. But well. Go Slowly. :)
One other point, if anyone wants to move over let them, and help them do it. If they are choosing to switch they could be very very helpful down the road.
Fedora is a bleeding-edge distro with a rapid release cycle and relatively short support period. If Linux makes sense for you at all, you should probably be looking at Ubuntu LTS or Debian on the desktop, and RHEL/CENTOS/Debian for servers. Fedora would not be my first (or even second...) choice for deployment in an enterprise environment, unless most of your users are *NIX software developers (and they're developing for RHEL/CENTOS as the target environment).
I've been there, done that, and gotten the pink slip. No, not literally - but I've looked into doing things like this in the past.
Consider for a second why you want to do this before you approach it, as well as the added overhead of maintaining multiple, divergent systems.
As for Exchange -> Gmail... why? Seems like a (significant) downgrade to me, and I'm particularly un-fond of Exchange.
If you're considering multiple apps under WINE and completely abandon the existing OS, I suspect you're a bit of a fanatic (or simply inexperienced). You want to do something like this with baby steps. One application at a time!
What's the justification? Licensing costs? Avoiding malware? Reducing management overhead? What is your end goal?
The only conceivable time I can imagine moving common workstations to LInux right now is if you're running on ancient XP machines and/or the necessary applications are either minimal and do not necessarily require Windows, or you plan to move to something like XenApp for important Windows apps. Moving already-licensed W7 machines to Linux "just because" seems stupid unless there's a good time/money management reason for it.
IF you're silly enough to approach this, I suggest you look at user requirements - and then start replacing and/or migrating one thing at a time. If you want to get rid of Exchange, I suggest you look at that first, consider options, and do a migration only once you've figured out that it makes sense after considering all use-case scenarios.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
Your job as the IT resource for the organization is to give the staff the tools that they need to do their job. Do the sales people want new tools, or are you trying to force new tools upon them? The sales staff pays your salary. As much as it sucks to hear it, that is the bottom line. They have a workflow and a way of doing things that is centered on the tools they have. Why are you trying to upset the apple cart?
Linux has matured to the point where if you are starting from scratch, it is a viable path to take. You can get the functionality you need at a fraction of the cost. Linux is not enough better than Windows (or OSX) to migrate onto it (for most organizations). If you like Linux, bring it in where you can. If you need to develop a new application, consider a LAMP stack instead of SQL and IIS. If your boss randomly starts whining about licensing costs for Office, suggest OpenOffice.
Do not take it upon yourself to "make things better" if you are the only person who seems to care. Let the users tell you what they need, and help guide them to the best solution. I have seen careers ruined by people who truly wanted to make things better, but were too caught up in their own heads to realize that nobody else seemed to care. They end up "solving" problems that do not need to be solved, and in the process create a lot of upset and headaches. Migrations are never simple. Often times going from one version of an application to another is a big enough headache, nevermind one OS to another.
If you put a wolverine, a badger, and a mountain lion into a box, will they cuddle?
You haven't provided anywhere near enough information to give useful advice. What are you trying to accomplish? What are the users doing? What tools are they using (releases count), etc. Who would be using Linux and why (if it's going to be low cost windows replacements, then perhaps rehink your choice of distribution...)
You need to trade off budget, vs. requirements vs. desiderata .. it's why IT is a profession not a hobby ;>
As to the question you asked, if you keep things on Exchange, and CIFS everyone can share. If you migrate to IMAP based servers everyone can share, except for calendaring (outlook's Calendar features are not the same as what you get with Google Apps, so be careful what you threaten your user community with).
How do Sales and Marketing communicate? What do they need to collaborate on? If it's just PDF documents from Marketing->Sales then the question is pretty meaningless. If they need to coauthor documents you might have very different Requirements.
Personally I work in a mixed Windows/Linux environment, and sometimes use personal Macs attached. Engineering is CentOS based, my Linux laptop is Ubuntu, my Windows laptop is XP and my Windows VM inside of the Ubuntu environment is Win7u. Macs are aged PPC based devices.
Depending on just what you are trying to share and WHY makes all the difference ... but it can be done. Trivially in many cases; less so in others.
As others aptly noted, taking Excel away from power users is seldom a successful strategy.
Company trade secrets, financial information, etc should *never* be hosted on a 3rd party site. Emails, right or wrong, will have that information...or at least internal emails will. Of course, once you go to gmail there's no such thing as internal email.
I see this general idea posted a lot, but in actual fact real corporations and governments frequently trust such information to third parties. Contractors and subcontractors are privy not only to the government secrets that they are working with to perform their duties, but each other's internal documents. Companies like Iron Mountain based their entire business model on archiving, protecting, and, under the proper conditions, destroying other company's internal documents.
The Fortune 50 company I used to work for contracted their entire corporate IT infrastructure to Dell. Dell provided workstations, IT help desk, and ran all the internal and external servers. Below the level of the CTO pretty much every person in the IT department actually worked for Dell.
Security companies like Brinks provide all the physical security including guards and cameras for lots of companies. The guards who work for our security contractor have more access to our building than I do as a regular employee.
In short, most companies of any size already trust a good portion of their internal information to other companies on a regular and ongoing basis. How is this different? You write the contract to ensure severe penalties for the third party in the event the information is deliberately compromised, less serve penalties for accidental compromise and you do business.
I don't need a million points of light, just two points of multi-mode fiber and a 10 Gig-E router.
What nonsense!
Personally, I think an Outlook/Exchange solution is much more productive for heavy office email users than the clunky thin-client Gmail offers, but this is one of the most egregious examples of FUD-seeding I've seen.
Java, C#, C++, SQL, Oracle, SQL Server, Perl, VBA, .NET, Visual Studio,....Linux and Windows
I got that from a job posting that my father-in-law sent me.
And at 35, you're oooooolllllldddd in corporate IT.
RIP America
July 4, 1776 - September 11, 2001
That will teach you to question the modding system!
A friend of mine tried this with her rather savvy users, but the churn in Fedora created too much work to keep up with. It worked fine, but they ended up switching to Ubuntu LTS for the longer support lifetime, since CentOS 5 was getting a little old. If you prefer the Fedora ecosystem, RHEL 6 was just released, and CentOS 6 will be out soon.
There's no failure quite as dissatisfying as a complete and total solution to the wrong problem.