00:53 <jfred> by any chance is it possible to manually enable outbound network access to a grain temporarily that wasn't designed to request it?

00:54 <jfred> for context, I'm trying to import a wordpress media library from an older non-sandstorm wordpress instance before a migration, and apparently that involves downloading each file from its original source

00:56 <jfred> (I thought it might make a powerbox request for a connection to the source host when attempting to do so, but it doesn't seem to)

01:03 <JacobWeisz[m]> Not really, no. Sandstorm sandboxes don't block network access so much as they simply don't support it.

01:23 <jfred> Ah, hmm

01:24 <isd> (From the app's perspective, it's running on a box that doesn't have any nics).

01:35 <jfred> okay, another possible solution: presumably the grain's filesystem is somewhere on disk, could I access that directly from the sandstorm host and drop my files into it? (not sure if it's a mounted block device or just a directory or something)

01:37 <JacobWeisz[m]> I think that should work.

01:37 <JacobWeisz[m]> Grains are stored in some subdirectory of /opt/sandstorm.

01:38 <jfred> aha yup found it - /opt/sandstorm/var/sandstorm/grains/<id>/sandbox

01:40 <jfred> now, the wordpress functionality that lets you add media from the local filesystem to the WP database seems to be missing though (possibly disabled as part of the sandstorm package, I suppose)

01:41 <jfred> oh, never mind, that's a plugin

03:35 <jfred> yup, to confirm, copying my files right into my wordpress grain's sandbox dir seems to work just fine

03:37 <jfred> heh, this grain is getting very large... I wonder if sandstorm would balk at trying to restore a >600 MB grain backup

03:59 <jfred> (it also takes a very long time to republish the site, but it looks like that's something I'll either have to live with or start working on the wordpress sandstorm package myself to fix :P)

04:09 <JacobWeisz[m]> Sandstorm can work with 600 MB grains pretty well, I imagine the backup might shrink too a bit because it's compressed, depending on what's in it.

04:10 <JacobWeisz[m]> The big thing you are presumably hitting is that publish is turning your blog into a static site. :)

12:37 <JacobWeisz[m]> Some much needed GitHub Sponsors features: https://github.blog/changelog/2021-04-06-custom-amounts-and-one-time-payments-rolling-out-to-github-sponsors/

13:51 <jfred> JacobWeisz[m]: yeah, I just wasn't expecting it to take quite as long as it does to build it as a static site haha

13:51 <jfred> (~10m I think?)

22:09 <jfred> so, this is a bit concerning... I have a wordpress instance published as a static site through sandstorm, but recently the nginx reverse proxy I have in front of it has started returning 502s sometimes until I restart sandstorm

22:10 <jfred> huh, and in the grain through sandstorm as well

22:10 <jfred> only thing in the debug log is: ...*** couldn't connect to grain (Error: remote exception: stream disconnected prematurely) ***

22:15 <jfred> (this is not the grain I've been mucking around with the filesystem on, fwiw)

22:17 <jfred> ...oh

22:17 <jfred> http://paste.debian.net/1192678/

22:18 <jfred> filesystem looks to have plenty of space left, are there per-grain quotas or something?

22:22 <JacobWeisz[m]> Shouldn't be.

22:23 <JacobWeisz[m]> The static hosting area is weird though, I wonder if its specific to that.

22:34 <jfred> mmm, turns out that "no space left on device" is a bit misleading and it was https://askubuntu.com/a/1088275

22:36 <jfred> if something's trying to inotify watch everything in that grain's sandbox directory, it has... a lot of files right now

23:19 <isd> Yeah, that'd be the supervisor trying to keep track of how much storage a grain is using. Could you open an issue? Would be nice to handle this more gracefully.