Bryan's picture
I am a new immunology professor/scientist, and am in the process of setting up my first lab. I've used ubuntu as a desktop system for 5+ years, managed a small ubuntu-based cluster for the past 3 (which I did not setup, btw), and as such am fairly familiar with linux/ubuntu as a desktop environment.  I'm also semi-familier with netowrking and CIFS/SAMBA as well.

As part of my lab setup process I am putting in place a minimal server. The main purpose is file storage (we'll generate a few TB of data every year), but I also want a wiki (for experimental protocols) and backup capacity. because there server will be lightly used (file storage/access being the major thing; with less than a dozen users), I was planning on using modest computer running ubuntu as a front-end, and rack-mounted RAID for expandable storage (i.e. devices like these):

The server will be managed primarily (hopefully solely) via remote desktop.

My goal for the server is fairly minimal; I want to:
1) Run a file server, accessed internally via SAMBA/CIFS
2) Provide external access to these files via SFTP or another protocol
3) Run a Wiki
4) Backup lab computers nightly to the server
5) Mirror the server nightly onto a secondary (backup) server

The actual amount of traffic will be quite small (a dozen or so users, maximum, accessing from a handful of computers). I have been looking at some options, and turnkey linux seems to be an easy way to get these things up and running. That said, having never used turnkey before I am not certain this is the route I want to go.

So now for my questions:
1) Is turnkey the easiest way to implement this? If not, what would you recommend?
2) From the instilation sections of this webpage it appears that turnkey is most often installed as a virtual machine.  While I'm very familier with the use of virtualbox, I was wondering if this were truely nessisary?  Can I not reduce the computing overhead by simply installing turnkey directly onto hardware?  This seems the logical course; especially considering the minimal server capacities I require.  AFAIK, the x86.iso should install like any other distribution.
3) If #2 is possible, how do I "add" another turnkey module to the installed OS (i.e. if I install the File Server Appliance, how do I then add the mediawiki package)?  Can the second appliance be added to the installed appliance, or do I have to virtualize one?

Any other advice is appreciated



Alon Swartz's picture

I think the easiest way to get up and running is to use TurnKey Fileserver. You can use the ISO to install onto bare-metal (ie. real hardware). This will provide you with 1 and 2 of your goals in a couple of minutes.

As for goal 4, thats up to you - whatever your preference is to transfer backups to the server (rsync?).

As for goal 5, I'd recommend you take a look at TurnKey Backup and Migration (TKLBAM) for offsite encrypted backups to Amazon S3. Just initialize TKLBAM and it will take care of the rest (but always test your backups).

Now for goal 3 - You will either need to install the wiki yourself, or use another machine and install one of the wiki appliances. If you have a spare machine lying around (even a low-end one), that would probably be easiest.

Hope the above helps

Bryan's picture

Thanx Alon,

For legal reasons we cannot use a cloud for storage; we have patient and other confidential data that is illegal for us to store off-site (even if encrypted).  A fellow researcher is setting up a simular server in his lab, and I will be backing my server onto his (in exchange, he's backing up his server onto mine).

I'm glad to hear that getting the file server up and running on bare metal will be easy.  I'd prefer to not add a second computer for the wiki (simply for space reasons), so I'll have to instal the wiki system myself (or maybe I'll just virtualize it).

Thank you again,


Jeremy Davis's picture

Although it may take a little patience and trial and error, as by default the FIleserver uses the standard port 80 for eXtplorer (a web fileserver front end - which you may find redundant anyway in your setup).

Under the hood TKL is Ubuntu 10.04 so it'd just be a case of deciding which wiki you like (I'd be inclined to go for something that is in the Ubuntu repos if possible). You could still use TKLBAM for your offsite backups, but you'll need to set up a custom cron job (so you specify the remote location rather than use S3). You miss out on the ease of use that you get with the Hub but it should fulfill your needs. Info about setting up TKLBAM manually is in the docs (although not specifically on setting up a cron job for it, you'll have to join the dots there).

Add new comment