Linux - NetworkingThis forum is for any issue related to networks or networking.
Routing, network cards, OSI, etc. Anything is fair game.
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
SDN 101: An Introduction to Software Defined Networking
Discover the advantages of SDN.
SDN has quickly become one of the hottest trends in IT. But not all SDN solutions offer real software-defined functionality. As more enterprises consider SDN, they want to know, “What is SDN? And what are the real benefits?” If you're ready to explore the advantages of SDN, and want to know how it should be implemented within your enterprise, start by reading our introductory white paper.
Click Here to receive this Complete Guide absolutely free.
I know that you'll want a lot of RAM, and drive space, but i don't think that cpu speed is that much of a deal. The RAM will just be able to hold all and any instances of the programs that are running, and to support many connections. As long as you have a lot of ram, disk access won't be happening that much, except for and saving files... (once they're loaded they'll stay in mem)
Is this for a business or for a home office? What I've written below mostly concerns the needs of a small business... if it's just a home office, you might not care about backups so much, for example.
1) It depends on what you'll be doing with those machines; what software they will be running. For basic Office software (word processing, spreadsheets, web browsing, email etc), you can get by with the smallest available drives and roughly 256MB. RAM is pretty cheap though, so you might as well make your users comfortable - give them 512 unless you're on a limited budget. You want to actively discourage using local storage for these machines, which is why I say the smallest drives available. Generally I can't find anything less than a 40GB drive new, but the most you should need is around 10GB. You could try to set up some sort of local caching system to improve speed and reduce network traffic, but with only three clients it's probably not worth the hassle.
2) Your two primary concerns for a basic file server are storage and backups. I cannot stress enough that these two concerns are basically of equal priority. If you need reliability as well, spend less on the CPU and get a decent RAID setup. Hardware RAID is better performing and more reliable, but obviously more expensive. Plan for at least the next three years (storage requirements will increase over time, and usually not in a linear fashion), and immediately start thinking about how you will back it up. Don't even purchase the server until you've planned for the backup. Tape is expensive, but fairly reliable, otherwise you can try burning DVDs. Never done that myself for an office, but for three users it should be acceptable.
3) For serving files, a single P2 700 would be fine. Hell, a Pentium 133 could probably do the job... most of the time, that processor is going to be idle, unless you start adding different services. If you're planning to make it an intranet server or do other processing with it in the near future, then plan for that, but if you're only going to serve files for the next five years, don't waste money on a dual CPU setup - it's money down the drain that could have gone towards better storage/backup hardware. People used to run excellent Novell servers on 386 machines, even Pentium class clients. Your biggest potential bottlenecks are disk and network bandwidth, not CPU speed. If you get a dual CPU file server, I can almost guarantee you'll see 0% usage of that second CPU.
Definately some information on what specifically the workstations will be doing would be helpful. If you are working with massive video files, your needs change compared to, say, an email exchange type setup. For a "general" recommendation, I strongly support CroMagnon.
RAID system can be very cheap to setup. I have a samba server at home that has a RAID 1 setup, the RAID card cost me $65 and has 2x 120GB 7200 rpm Seagate drives. Runs very well for the 4 machines that I have hanging of it, all though they're all not connected at the same time, but I only use it for saving files so it doesn't do a lot of work. But should one drive fail, at least the data is safe, then I just need to buy another one.
Thanks for the replys, it really does help a lot! This computer is going to be for a small business and will run mostly office thing, word, gimp and so on. I would however like to be able to get some video going and music as it's a very easy going business. I have no clue what the heck RAID is honestly. You can only be smart or good with so many things and I've just run out, computers, while I have a good understanding Linux is new to me (4 months) and networking is very new as this is the first I will have set up.
1) A server is basicly just an easy way to share a HD with the workstations correct?
2) Is there a "most stable" hardware?
Thanks again and any links to how to set up a Linux network, distros to use or anything else will be worth it's weight in gold to me!
RAID = redundant array of independant drives. That part I could help with. Clueless on the rest though. I just know that there are different kinds of RAID. One kind mirrors data for redundancy, prevent lose of data in case of drive failure. Another kind make the system faster by spreading the data out. Can make a big difference from what I have read.
Hope someone can help with the rest. Here is a link that may help explain the RAID differences.
I can add a little to that. RAID 0 utilizes an even number of equally sized drives, and any data sent to the RAID controller (which handles the RAID setup and the drives) will split data across the disks. RAID 0 can be thought of the opposite of patitioning (partioning splits a disk, and RAID 0 adds them together as one logical drive - you only see one drive). The benefit with this is that you get to double or triple or quadruple the amount of space a drive can have. The downside with RAID 0 is that if one drive fails, you lose all data because it is split up.
RAID 1 uses mirroring. If you have two disks on RAID 1, the same data is written to both disks, making a second copy of it. The benefit of this is that you get faster hard drive access (one disk will access the first block, while the second will move ahead and load the following block, etc). However, with this setup you do not gain as much space; it is split in two.
RAID 10 i think.. or 1+0 (i'm not quite sure) utilizes both. You can have two sets of two disks like a tree.
| | | |
The bottom left two disks and the bottom right two disks would be in a RAID 0 to combine them into two drives (the two points at the top) and then RAID 1 would split the data across those two. You can also have it in reverse where RAID 1 is at the bottom and RAID 0 at the top.
I'm not sure about the rest (4, 6 i don't know the other numbers), and i know a little bit about RAID 5. RAID 5 with three equally sized disks, for example, not only splits data across the drives, but also stores tiny extra information. This information is used incase one of your drives fails, you replace it with an equally sized disk into the array, and issue a command. This command will utilize the extra information bits stored on each of the other disks to rebuild the lost information back onto the empty drive, thus circumventing data loss. This won't work if you have two or more drives fail simultaneously. Your total disk space becomes 2/3 of the total with RAID 5.
In my opinion RAID 5 would be the best for a small setup, however take note that each RAID configuration has requirements on the number of disks you may use, and how you may set them up.
Hope this was clear enough to give you a basic understanding..
1) A server is just another computer, devoted to automated tasks. Those tasks can be whatever you imagine a computer can do without human operation. Sharing hard drives is usually called "file serving". If you are setting up Windows clients, you will want to learn about Samba to set up a file server. You will probably need to do a lot of reading and searching to get this set up exactly right, but once it's running properly it will usually stay that way for years. This reminds me of an important point - write everything down. What I mean is, as you're configuring the server, note down every single step you take to make the system work. Any problems you fix, write those down too. Put it in a text file if you want, but print it out and store it in a filing cabinet or something. Once things are working, you will forget what you've done, and if something dies in three years time, you will thank me for that document, because it won't take you a whole week of angry users to set the thing back up again.
2) This is a partially religious question. Most hardware is pretty good these days... if you build it yourself, I personally like Asus hardware, and have never had a problem with Maxtor or IBM drives, but this is pretty subjective. I can almost guarantee that after writing this, someone will come along and say "I had three Maxtor drives fail within a week, they are crap!". What you should really care about is the warranty... with a three year warranty, even if the drives fail every twelve months you will get a replacement for minimal expense and get three years for your money (of course, if the drives regularly fail every twelve months, start looking for a new manufacturer). This is a 'bad-luck' scenario though - most hard drives will keep running for five years or more, depending on the workload. Did I mention the importance of backups?
RAID tutorial - explains the various RAID levels and their pros and cons.
Re: Wireless router - yes, this is almost a certainty. Just be aware that there are two types of cable, one of which won't work depending on the router (crossover and passthrough cables). If the router was built to plug in to any normal port on a hub, you will likely need a crossover cable to connect it to your PC. If it was built to plug in to an uplink port on a hub, then a passthrough cable should work.
Holly crap I'm in over my head, I'm going to go to a computer store and find somebody A LOT smarter them myself to help with this. At least you all healped point me in the right direction, educate myself and learn what to ask. Thank you all very much and I'm going to go tomorrow and look into this more. This system is my a business I'm just starting and I thought I could deal with the server and all on my own. In most cases I could but I don't have the time to learn all I need to know so THANK YOU ALL AGAIN for you time and replys!