Raspberry Pi Server - Other Alternatives.

I have a Pi B and Pi2 B.  The Pi B ran into performance issues with Owncloud but the 2 runs it well.  I run Owncloud on the 2 with a dynamic DNS client and web exposed and we have 4 phones, two tablets and 6 computers syncing with it.  The Pi runs my internal server with a 1TB drive.  It holds all our other documents, pictures, music, etc that's not on Owncloud.  It's the older stuff that I don't keep synced.  It's also the backup server for the Owncloud pi2.  The internal shares all its data via samba for the windows clients and NFS for the Linux clients and minidlna allows standalone devices like a blueray player to access the media as well.

The orange Pi, period, beaglebone, or small x86 systems such as miniityx boards are all alternatives but the Pi2 works well.

I have a Pi B and Pi2 B.  The Pi B ran into performance issues with Owncloud but the 2 runs it well.

This is just the point. To make a project is more than acceptable to choose the better performances / prices hardware; but if the project should really work for a long time, maybe the worth to invest few more $ and buy a more reliable than super-cheap device anyway.

To make a server IMHO it is different, especially when the server should serve This is the reason that in cases like this I wait until there is not the right budget (or I induce a client) and make the better choice respect the absolute cheaper.

Enrico

On the other hand, for someone gaining experience, first getting the cheapest option is a good way to learn "what is needed".

Example: I had never owned a tablet. What use would I have for one??? Got one of the cheapest, found out what the thing is useful for, and that I need a bit more performance. You can argue that maybe I should have gotten the more performant one immediately. But that's not true: I didn't know what "extra performance" I needed. Better screen resolution? More RAM? More storage? Back then I didn't know, now I do.

Similar case here. Getting the cheapest option first will allow you to learn where that option is lacking **in your case**. Of course if from the outset you can predict it won't satisfy the intended use-case, then it's no use. But if the use-case is vague, unpredictable or may expand in the future, then "maybe this will suffice" and getting the cheapest option is a good way to gain experience in figuring out what you need in the end.

On the other hand, for someone gaining experience, first getting the cheapest option is a good way to learn "what is needed".

Example: I had never owned a tablet. What use would I have for one??? Got one of the cheapest, found out what the thing is useful for, and that I need a bit more performance. You can argue that maybe I should have gotten the more performant one immediately. But that's not true: I didn't know what "extra performance" I needed. Better screen resolution? More RAM? More storage? Back then I didn't know, now I do.

Similar case here. Getting the cheapest option first will allow you to learn where that option is lacking **in your case**. Of course if from the outset you can predict it won't satisfy the intended use-case, then it's no use. But if the use-case is vague, unpredictable or may expand in the future, then "maybe this will suffice" and getting the cheapest option is a good way to gain experience in figuring out what you need in the end.

Almost correct Wolff, but first of all this is not an "unpredictable case" as the user explicitly asked suggestions on making a server.

IMHO there are some braking lines where you can't go down in every case. Especially for newbies and there is a reason, supposing that the newbie is not lazying trying just for fun and just because he has nothing better to do (no good films, no friends to go with to the pub, bored of bowling ); when a guy, a very young guy max 6-7 years ask to his parents to play guitar the most probable thing they do if decide to buy a guitar to him is to find a crap just-few-more-than-a-toy guitar. He start playing, study and discover that guitar has a so bad sound that it's no worth. But is the guitar playing bad also when the kid has a sufficient knowledge to play something well. The same is the tablet you never owned. sure it is not the worth you buy a super device investing 5-600$ but if you just try with a Android 2.3 25$ tablets it is very probable that you abandon because you are so far from what a tablet toady is that you continue without any idea.

There is also another aspect, involving the budget. Suppose that you decide to make a server and buy a 177$ Gizmo2 or 79$ SBC like the last XU4. Then you add .... 40$ for a fast 500 Mb HDD USB3. You have spent an amount between 200 and 250$. You have a server, that is what you want, reliable and with good performances. Then you come happy and buy a second HDD and invest another 40$. You have again a server, based on a good SBC and better. And so on.

Instead it is not rare that trying with several low cost devices, adding components, making tests with a platform not useful to reach the goal the expenses count at the end of the experiments result higher than what expected.

Enrico

Agreed. Again: If you can easily predict that the "cheapest option" will be lacking in some respect, then you should immediately get the more advanced/expensive option.

You're saying "server". You have an image with "server" that may not be the same as what the thread-starter wants/needs.

If "server" means: "a system that can download movies in the background and then serve them on-demand to a media player on my network", then a $35 raspberry pi will suffice. Sure the raspberry pi has lousy disk-performance. But does that hinder the background downloading? No! Does it influence the media streaming at all? No! Of course the disk and ethernet performance come into play if you want to move big files between your PC and the "server", but if you never do that, why spend the extra money on that?

Often you can get a $35 option and "twice as much" for a little extra, say $40. Then even if you don't know what it would be useful for, you might chose to get the bigger option. Now in this case the choice seems to be a $35 raspberry pi or a $70 XU4. You might think that this is quite a jump. But if you count all the extras, SD card, powersupply, harddisks the difference is much less than 50%. The small price-increase might be acceptable for some people even if they don't really know right now if it will be useful, but for others the difference is significant.

You're saying "server". You have an image with "server" that may not be the same as what the thread-starter wants/needs.

This maybe, I have considered this possibility but why not use the correct terms for the things?

Anyway, in the case I have to build a media server (that is lightweight than a traditional server for sure), I should provide at least:

• Good transfer speed because good videos and films today are streamed in MKV format that occupy some giga of space. Try to transfer on a 10/100 lan or with an external USB 2 HDD just only 10 films to put them on your media server ...
• A reasonable external storage space, half or better 1 Tb just for the same reason, films and music
• 90% of the users I know  expect that their media server is permanently downloading torrents, also when they are watching videos.
• There is a good percentage of people that like media servers because gives the possibility to access different sources from different PCs.

And more. Despite that there is a specific distribution for the Raspberry PI to make a good quality media server. With all the hardware imposed limitis.

Enrico