Making RockPro64 a NetBSD Server | Hacker News Reader

Making RockPro64 a NetBSD Server

blog.netbsd.org

83 pointstodsacerdoti16 comments

Have netbsd head running on 4 nas cases on this board with zfs. It's been very solid.

Do you use some specific distribution with support for NAS features or is it just a regular netbsd which you set up by yourself

zfs as a filesystem and nfs, samba or iscsi are enough for that usecase.

This is really neat, although I tried for 6+ months to get the SBC, NAS case, and other accessories before I finally gave up. Something was always out of stock (usually the NAS case). I'm sure my timing was bad, but I didn't care enough to write a watcher bot.

Has availability improved with Pine64 stuff?

I ordered the nas case and a rock64 pro about 2 weeks ago? And it came in recently so imagine some things are in stock.

I've been running my rockpro for a few years now, it's been awesome.

https://www.pine64.org/availability-and-shipping-status/ is the canonical source for what's in stock and when it'll ship.

Got an H64 a few months ago as an alternative to a RPi. Definitely much easier to get the H64 than a pi.

Bit disapointed by the support in general though. I expected the Pine stuff to be better than "random boards from no name companies on Aliexpress" in terms of HW documentation and software support, but it wasn't that great. Software support is totally left in the hands of the community, and I couldn't get the spec for the uFL-like on-board antenna connector anywhere, total silence on their forums for months.

I mean, it works well overall (running Armbian), but I also had random aliexpress boards that had to be fully reverse engineered work pretty well.

Its the pandemic. I got my NAS before without a problem. Currently waiting for the tab and laptop to stock up again.

I lack the background to understand why big endian patches are required. They're ARM Cortex chips, they support little endian, right? Why would you use big endian?

The world isn't simply filled with people who are content to do the same thing as everyone else. For some, it's purely aesthetic. For others, it's preferred to have an environment where software that makes bad assumptions won't work. For many, it's at least partly just because we can.

Not everyone does things just because everyone else does them. If that were the case, we'd all be running Windows. Some people do things because it's what everyone else isn't doing.

That doesn't really answer the question. But the philosophical puff piece is 'appreciated'

I thought of it as the voiceover to a Super Bowl commercial — traditionally car commercial, but these days it would be for crypto.

Software that makes bad assumptions tends to fail on in a million times on 'normal' systems, but everytime on others. Driving those one in a million cases out makes everything a lot more reliable for us all.

Sometimes people use oddball architectures specifically for the purpose of exercising the portability considerations in the source and/or build process.

It tends to uncover bugs otherwise unnoticed, or just shortcomings in the abstractions.

There doesn't need to be any sort of runtime benefit to make it a useful exercise.

I hadn't quite caught that this was replacing a SunBlade 2500. I get it now -- he bought this board specifically to work on big endian support, and was already doing that work on the SunBlade. I had misunderstood this to be someone getting a board for some other purpose and just choosing to run in big endian mode for fun.

Reply on news.ycombinator.com