21:29 <anthropy> hey guys :) I've been trying to upload an old backup of a dokuwiki to sandstorm (which I salvaged from an old broken sandstorm-- but couldn't do an export there as it didn't start so don't have a proper backup, only the filesystem), but I've ran into an issue with my method; I've tried to make a backup of a fresh dokuwiki grain, modifying that zip file by putting the desired dokuwiki files in there, and uploading that, however that

21:29 <anthropy> doesn't work as the backup files seem to have a hash in their metadata which it verifies after upload. Long story short, I couldn't find a way to get around that signature check, is there a way I can turn it off temporarily?

21:36 <kentonv> anthropy, there's no hash in the metadata. The structure is declared here: https://github.com/sandstorm-io/sandstorm/blob/master/src/sandstorm/grain.capnp#L614

21:36 <kentonv> it should work to swap out files

21:36 <kentonv> what error are you seeing?

21:37 <anthropy> "invalid backup file [500]"

21:37 <anthropy> I'm specifically modifying the /lib/dokuwiki/ directory, the rest is as the grain says

21:38 <anthropy> also thanks for looking into it ^^ appreciate it

21:38 <kentonv> does /opt/sandstorm/var/log/sandstorm.log have any more error info?

21:39 <anthropy> let me check

21:40 <kentonv> yeah there should be a line that starts with "Unzip failure:"

21:40 <kentonv> and then more info

21:41 <anthropy> yea I tailed it, had to redo it to get a clean output :) one sec lemme pastebin it

21:41 <anthropy> https://pastebin.com/hYtSZS7d

21:42 <anthropy> wait, I'm gonna try see if maybe it's putting a folder inside the zip

21:42 <kentonv> yeah that could be it

21:43 <kentonv> it looks like it can't find the top-level file "metadata" and directory "data"

21:43 <anthropy> yea I noticed that too, I hadn't checked the logs before tbh

21:45 <anthropy> squee it loaded! and everything is still there C: great! thanks for the pointer, I should've checked the logs first lol

21:45 <kentonv> np, it's often not really obvious that the logs would contain more info

21:58 <anthropy> while i'm here anyway, I've as long as i can remember had trouble making new wekan grains somehow once every while. I've never been able to smash the issue despite having had multiple sandstorm servers with different distros to host it on. Has anyone ever had issues like that with wekan grains? here's some logs: https://pastebin.com/NVFZqWMT the hardware has ECC memory and xeon CPU and BTRFS with weekly scrubs .. somehow the grains

21:58 <anthropy> occasionally start working after a long while of leaving them open? and existing grains still work, and i can even clone those to create new grains as emergency.

22:00 <kentonv> hmm, try downloading a backup and look at the mongo.log inside, it looks like mongo is not starting for some reason

22:08 <anthropy> https://pastebin.com/QNbTnyER the logs inside there all look like this, but there's many of those logs, as if it's constantly restarted

22:10 <kentonv> odd, doesn't look like anything wrong.

22:11 <kentonv> could be running out of RAM I suppose

22:56 <anthropy> it has 8gb ram and doesn't seem to be running out.. but yea it's strange