On 19 August 2014 22:44, James
Savyasachee Jha
writes: ... I think your curiosity about the file systems Arch supports can be satisfied by this page: https://wiki.archlinux.org/index.php/File_systems
Thanks for the response, but, I'm not much for top posting, unless that is the norm for this list? (or you are mobile_phonic?).....
Some folks are careless, others can't avoid it sometimes; but that doesn't mean they represent the "norm" of anywhere.
Yes, just what I was looking for. It does not mention HDFS, but your package repository does show support for Hadoop. I believe that Hadoop must run on top of HDFS.....
I believe Hadoop includes HDFS, so you can expect full functionality by installing the package available in AUR (an unsupported repository of build scripts of varying quality). I see some documentation is in the wiki already. [1]
What I was really looking for is somebody on Arch that is running Mesos and Spark on top of BTRFS. Possible running hadoop on top of mesos. I see the dev "dseg" but is there a workgroup hacking away at mesos and the companion apps? I see "spark" in the packages but I do not see support for the scala programing language. I can hack my way through on these and many other packages related to cluster computing, but it would greatly help if there is a group of like minded folks on Arch? A (mesos)cluster lead?
I am not aware of such a group. Personally, I have been using a pre-built CentOS VM for some big data simulations I got into not too long ago, but haven't got around to the administration aspects of it. If you are looking for a particular feature from a package, and it's not available, get in touch with the maintainer to see if she is aware of this missing feature. Otherwise, DIY. If you need like-minded folks, or seek an expert on the topic, you might want to contact the users who are maintaining these packages (including 'dseg').
So I'm looking for guidance as to what/how to quickly pursue Mesos & spark & hadoop on top of Arch linux. That way I can build on top of btrfs later, unless that is what folks interested in distributed/cluster computing are already doing on Arch? (Arch) Birds of a feather on mesos et. al. ?
If you don't get any favourable responses to this call, it is up to you to, well, DIY. Tailor this distribution to suit your needs. If it all looks like too much work, look for a distribution that has it all ready to go. Arch Linux presently supports only i686 and x86_64, officially. Anything else, you look for a community effort. There is a project for classic ARM, which others have already pointed out. Otherwise, DIY, if you insist. [1] https://wiki.archlinux.org/index.php/Hadoop -- GPG/PGP ID: C0711BF1