bane

I've noticed that people seem to also have two working definitions of "reliability".

1. "This works as expected to without error or repair for longer than expected"

2. "This is easy and straightforward to repair, so when it breaks the halt in operations is only temporary"

I don't know why people seem to use #2 when it is obviously not the meaning of reliability, but some version of it almost inevitably brought up when discussing an unreliable technology that somebody still prefers to use over the reliable one.

This thinking infests everything from computing to cars.

"Sure my PC crashes all the time, but its a simple reboot and some configuration tweaks to get it back online, takes like 30 seconds"

"I love my car, the service from the repair shop is white glove whenever I bring it in. I'm usually in and out in the same day"

"My stand mixer is amazing, I only have to 3d print a replacement for the central gear every 10 or 11 loaves of bread and it's a 30 second swap out to get it working again"

"I love my new clothes dryer, the maintenance contract gets somebody in within 24 hours to get it back online"

To the point of the post, very few organizations track long-term reliability and repairs for products, even major expensive ones like cars and appliances. The segment of consumer who really cares about this stuff just isn't large enough to make it really viable to do so, and the segment of consumers who follow definition #2 undercut them anyways.

People who name projects, please think very carefully if you want to use "Haystack" as the name of anything. There are literally thousands upon thousands of both overlapping and completely different products, projects, bands, initiatives, efforts, and so on with that name and all the possible variants you can think of (Haystax, Heystack, Heystax, Hay Stak, and so on).

And no, you probably won't be the first project with that name in whatever market/vertical/milieu that you are working in.

I found half a dozen different "Haystack" products and companies working in AI in 10 seconds of googling.

Please make it stop.

Happy Unraid user here.

The ability to mix and match drives in the main Unraid Array is of course the original feature and draw. Adding a few TB at a time for whatever leftover money I had after taxes each year is really appealing.

But they've added SSD write caching, VMs, Docker containers, a Docker "app store", and recently ZFS drive clusters (mostly for SSD storage).

It's pretty great and incredibly easy to admin. I presently have well over 125TB of mixed Unraid and ZFS cluster storage in a Fractal 7 XL. It's running around 30 containers, a handful of VMs, Tailscale and literally requires less than 20 minutes a week of system level administration (probably more like 5-10 minutes). Of course I'm spending far more than that messing with the actual apps, but that's a personal problem ;)

It gets regular updates, and I'm sure my uptime would exceed a couple of years except for reboots needed to handle the updates and the occasional power outage. You can ignore the updates of course to min-max your uptime. ZFS has been rock solid on my SSD array.

You can recreate the core array bits with a bit of effort and MergeFS and SnapRAID, add Docker, some VM host system, ZFS and a few other things and you can get Unraid "for free" with a fairly normal Linux distro, but the administrative overhead will be a bit more.

One tradeoff is that Unraid exposes a core set of features for each of these, but you could get to quite a bit more specific of a configuration if you go the regular Linux route. The Unraid devs are slowly adding more ZFS features (for example) to the regular interface, but it takes time. Some more expeditionary Unraid user attempt to use those features more or less at their risk with results reported in various forums.