Linux - ServerThis forum is for the discussion of Linux Software used in a server related context.
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.
Well here, let me describe what I'm looking for and we can take it from there.
I'm trying to build, as a proof of concept for now, a VERY lean linux server capable OS (a la CentOS/Debian) where the result would be used as the platform to build a storage solution, along the lines of NetApp.
I looked at OrangeJeos and that would have been an excellent starting point to then build that storage oriented OS I'm thinking of.
Please bear in mind that what I'm thinking of is not along the lines of OpenFiler or FreeNAS; its a headless server that could be used for enterprise scale storage solutions where HDDs plugged are automatically recognized and then RAID-ed accordingly (assuming some RAID schema is in place).
Actually, I'm basing it all on NetApp.
My two options are OpenBSD and CentOS, but somehow, and I have no concrete evidence of this, I feel that with the exception of ZFS, Linux generally has better handling of file systems. Again, I can easily be wrong in this regard.
If for nothing, I'd like to have a go at this and see if I can build this for my sake, for the sake of "I can". Call it a learning experience and if its stable and successful enough, it'll be open sourced.
Having said all that, what do you guys think my next step should be?
I was thinking of as follows:
Create a very LEAN CentOS
Work at the scripts that control the disks, including automatic "recognition" if a new one is plugged in
work on the CLI, possibly customizing a shell
release as alpha
Suggestions and constructive criticism are certainly welcome
In the end, Vanilla CentOS or OrangeJeos, it doesn't matter so long as the end result is achieved I suppose. But that still doesn't point me in the right direction. Are the links that I posted earlier appear to be a valid starting point?
A netinstall will give you a "seriously cut down version of CentOS" in less than 15 minutes, and then you can get started on:
2. Work at the scripts that control the disks, including automatic "recognition" if a new one is plugged in
because this step #2 seems to be the "heart" of your project. You will probably reach a wider audience by sharing your scripts with existing Red Hat/CentOS/Scientific admins, as the installed user base is well-entrenched.
Personally though, for your POC, I would start with a vanilla CentOS install (base install) and then specifically remove additional unneeded packages to make it even slimer. For example, I think even a base install has Bluetooth and cups which baffles me for a server - but i digress.
You could also recompile your own slim kernel, and then focus on building the scripts to utilize whatever you will be using for HDD management (mdraid???). i think that will be the meat-n-potatoes of your work. Once your POC is working, working towards sliming it down even further would be fairly easy to do.
(well snowpine said it before me .... so I agree :-) )