I have been an open-source junkie before it was cool. I haven’t used FreeNAS since around 2011 or so. I have been demoing it and googling again for a hard week, spending 8+ hours a day on R&D for my new venture. Some of the things are broken in the latest stable release like specific plugins and the pkg resources list (at least in the two on my install and burn to the ground installs). That said, back in the day when I was doing managed backup for my clients I’d install a custom-built Raid(x) *nix distro on commodity hardware in their location and pull out rsync backups to my remote (secured) server at my office. It worked like a champ and customers loved it. But with these weird problems I was facing today and over the week or so, and requiring my manual intercepting, I find myself wondering with using FreeNAS (VS TrueNas) if it was really scalable for my new venture.
The ZFS *BSD based file systems aren’t complicated and with a few basic bash script’s I’d write with cron doing the scheduling, would almost achieve the same results but without a shiny GUI WebUI. I suck at making things look pretty as a UI on web based platforms but it keeps bugging me that the issues are present in FreeNAS and maybe not so much in the paid TrueNAS. And I am sure if I roll my own, we would have intial growing pains too, but still how long has freenas been around?!. When I have to edit system standard files just to pkg update / pkg upgrade on a new stable release system, something just bugs me, and badly. It could have been me though, but I did RTFM and a lot.
So this got me to thinking; If I could create the complete back end using Bash/Perl scripts and hiring a UI dev to make it look shiny and new using my scripts and pulling data from them, could I make something better. Obviously I’d have to employ multiple people, so it wouldn’t necessarily be (free), but in my opinion when you pay for something robust you just expect it to work. (I know the OSS gods are striking down on me right now)
I know plenty of people love FreeNAS/TrueNAS but it has its quirks and stuff that bug me. Why do my clients have to rely on versioning done via client side apps? (Yes I get snapshots, great, but what a PITA when you are in charge of multiple locations/users) It should be easier and more streamlined. For instance to restore individual files (based upon a users request for what ever reason), I have to create a dataset, then create a share to contain that… oh nevermind, just one of my pet peeves on permissions and full DR etc.
Thoughts? I am sure I will get flammed for this, but Looking for input, etc.
@LTS_Tom – Love to pick you brain and brainstorm with ya on this