<JonTheNiceGuy[m]>
My sandstorm server won't start up, and I'm getting loads and loads of csr and json files in /opt/sandstorm/var/sandcats/https/jonspriggs.sandcats.io
mnutt_ has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
mnutt_ has joined #sandstorm
frigginglorious has joined #sandstorm
frigginglorious has quit [Quit: frigginglorious]
dwrensha has quit [Remote host closed the connection]
depb has joined #sandstorm
<depb>
Hi all, does anyone know if there are plans to port sandstorm to run on clusters with ARM processor?
dwrensha has joined #sandstorm
<asheesh>
Hi depb!
<asheesh>
It depends how much work you want to do, which is another way of saying mostly-no.
<asheesh>
There are currently no near-term plans to enable ARM for Sandstorm; we mostly think it'll make app QA harder, which isn't something we can commit resources to in order to make it worthwhile.
<asheesh>
Out of curiosity, what's your interest?
<asheesh>
I hear the request for ARM support periodically and I'd like to better-understand why people ask for it, if you're willing to share.
dwrensha has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0/20160726073904]]
dograt has quit [Ping timeout: 252 seconds]
<depb>
What a shame .... I ask it because I wanted equipment with Hardware Libre. Because now if all goes well it would be possible by EOMA68, each has A20 Dual-Core ARM Cortex A7, 1.2 GHz and 2GB of Ram, each of them. I would be in the clouds to have all FLOSSH XD And connect everything to Guifi.net it would be my dream come true XD
<maurer>
depb: So, from my understanding of the tech stacks involved, there's nothing fundamentally problematic in running on arm
<maurer>
depb: it's just that there's no releng for it
<maurer>
depb: so you'd have to do your own updates, builds, etc
<maurer>
and it might have bugs
<maurer>
but if you're already going the length to try to use exclusively free hardware, I suspect this might be within your risk/work tolerance
<maurer>
(you'd also have to do your own builds of spks, and wouldn't be able to use the app store)
dwrensha has joined #sandstorm
mnutt_ has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<mokomull>
A technical limitation that might prevent that is that the ARM SoC vendors aren't particularly well-known for supporting newer kernels after they're released, so the giant-blob-dump of $VENDOR may not ever support software interfaces required for Sandstorm.
<depb>
but in theory it would be possible, with possible bugs in the process ... of course? I simply compile all necessary software, like I do with soft from github?
<maurer>
depb: Yes, but as mokomull notes, you may want to test your kernel for the features sandstorm uses for sandboxing first before you go through the trouble
<mokomull>
I know several that are stuck at 3.4, and even on x86 Intel doesn't support anything other than a huge-patch atop 3.10 (with a bunch of useful features CONFIG_*=n'ed out) on the Edison.
<maurer>
depb: Unless you're building recent vanilla on your devices and having them work, it's kind of a crapshoot whether those features are enabled
dograt has joined #sandstorm
<mokomull>
depb: I also want to caution you that basically all of the vendors aren't know for quality or responsiveness to security issues, either. Libre may not
<mokomull>
Libre may not be quite the benefit you're imagining.
<zarvox>
There's also the additional issue that app packages are binary, so you'd also have to get app authors to build/release packages for ARM too. And we're not currently set up to deal with architectures that aren't x86_64.
todayman has joined #sandstorm
<zarvox>
Alternately, you'd have to automate the build pipeline enough that you could hand off some "source package" equivalent to automated builders. But then package signing gets complicated.
<zarvox>
Also, two of our dependencies, mongodb and meteor, didn't officially support running on ARM last I checked.
TC01_ has joined #sandstorm
<zarvox>
So, all that said, supporting Sandstorm on ARM isn't currently a priority for us. There's a lot of work for comparatively little gain, and there's a lot of unmitigatable risk from external factors.
<depb>
Ok I see, It will be difficult if not impossible. Thank you very much to all I'll have everything you have told me in mind. For now I will install sandstorm on x86-64 and I will try as an experiment on an ARM processor when I get my EOMA68
mnutt_ has joined #sandstorm
<mokomull>
zarvox: asheesh: I do have an AArch64 development board (the Dragonboard 410c) if y'all become interested in playing with it. WiFi was once shaky, but there's far newer drivers than I've had a chance to install.
TC01 has joined #sandstorm
amyers has quit [Ping timeout: 244 seconds]
<JonTheNiceGuy[m]>
Asheesh, would your access to my machine help?
mnutt_ has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<dwrensha>
nliao: it might be the meteor-spk does not yet support the version of meteor that you are using
amyers has joined #sandstorm
<nliao>
I'm using meteor version 1.4
Telesight has quit [Quit: Leaving.]
<dwrensha>
yeah, sadly that won't work yet. :(
<nliao>
thanks
<nliao>
will try down grading
pouledodue has joined #sandstorm
frigginglorious has joined #sandstorm
frigginglorious has quit [Client Quit]
mnutt_ has quit [Remote host closed the connection]
<xet7>
asheesh: We don't need to continue discuss terminology on that GitHub issue, I think changing terminology needs more thought and me reading more of the documentation for more complete proposal.
isd has joined #sandstorm
<asheesh>
xet7: OK!
<asheesh>
JonTheNiceGuy[m]: Yes! Just overwhelmed with other things at the moment.
<asheesh>
nliao: Glad that dwrensha was able to help you!
gambatte has quit [Ping timeout: 250 seconds]
prettyvanilla has quit [Quit: Konversation terminated!]
gambatte has joined #sandstorm
<mrdomino>
i am midway through writing skylark capnp -> c++ rules
<mrdomino>
hoo boy
<dwrensha>
skylark?
prettyvanilla has joined #sandstorm
amyers has quit [Ping timeout: 249 seconds]
nliao has quit [Quit: This computer has gone to sleep]
nliao has joined #sandstorm
amyers has joined #sandstorm
pouledodue has quit [Remote host closed the connection]