02:04 <mrdomino> if nothing else, this project has been a great intro to sandstorm's internals

08:40 <JonTheNiceGuy> TimMc: so, does that mean that if someone were to contribute a patch, they'd accept it? :)

13:00 <TimMc> JonTheNiceGuy: No idea. I could *imagine* them accepting a patch for something if it's well-tested but kept in a disabled state... but really no idea.

13:37 <yanivz> hi

13:38 <yanivz> i need to link rocket.chat to external website (embed) can someone tell how to this

13:42 <yanivz> anybody here?

14:09 <yanivz> help

16:13 <TonyTheLion> hello

16:13 <TonyTheLion> I wanted to know if anyone has tried to synchronize two sandstorm instances accross machines?

16:14 <TonyTheLion> as if you were running one instance and then wanted another back up instance with the same apps/data as the first instance

16:28 <TimMc> TonyTheLion: You could certainly do a periodic sync, but I don't think there's anything like a continuous sync.

16:28 <TimMc> I take nightly backups by taking a ZFS snapshot, backing up the snapshot, and then releasing it -- but instead of a backup, you could do an rsync over to another machine, and more frequently.

16:31 <TimMc> For reference, here's my snapshot-backup script: https://github.com/timmc/commapps/blob/master/scripts/backup/run-backup.sh

16:33 <TonyTheLion> I set it up with syncthing only sync'ing the /var directory so as to not overwrite the sandstorm config's wich are different on each machien

16:33 <TonyTheLion> I set it up with syncthing only sync'ing the /var directory so as to not overwrite the sandstorm config's wich are different on each machine

16:33 <TonyTheLion> but it doesn't seem to work

16:33 <TonyTheLion> it sync's the files

16:33 <TonyTheLion> but its like sandstorm doesn't care

16:34 <TonyTheLion> cause you can't tell in the UI that things were changed

16:57 <TimMc> Does that include the DB?

16:58 <TimMc> (looks like it does)

17:03 <TimMc> I'd be concerned about not having snapshotted syncing. It also sounds like you're trying to sync the database's backing files while the second Sandstorm is running?

17:39 <TonyTheLion> ye

17:39 <TonyTheLion> I am

17:39 <TonyTheLion> but that seems to mess up the db

17:40 <TonyTheLion> my second sandstorm instance's mongodb got corrupted

17:40 <TonyTheLion> :(

17:43 <TimMc> yup

17:44 <TimMc> What's your use-case?

17:44 <TonyTheLion> have another instance of the same sandstorm (with all user data) running as a redundancy

17:44 <TonyTheLion> if the first goes down, it can be seamlessly switched to the other

17:45 <TonyTheLion> I'm guessing thats not going to be as easy as it sounds

17:52 <TonyTheLion> I was thinking maybe i can set up mongodb replication

17:52 <TonyTheLion> for sandstorm's internal db

19:35 <TimMc> I suspect even with that it couldn't be seamless -- applications would still have their own internal state that they manage in arbitrary ways, and running applications would have to be relaunched and would lose in-memory state...

19:37 <TimMc> You could, however, sync snapshots over and only start the second sandstorm when needed. You'd lose the last few minutes of work, though.

19:39 <TimMc> It's possible that a reverse proxy sitting in front of both could buffer requests during the switchover, since sandstorm starts quickly... but all your websockets would have to be restarted, so I'm not sure there's any point in that.

22:54 <TonyTheLion> hello

22:54 <TonyTheLion> anyone know how I can connect to the sandstorm mongo db?

23:38 <TMM> hi all! question about capnproto, is there some documentation on the 'non ez' rpc interface? I'm trying to figure out how to make a multi-threaded dispatcher

23:39 <TMM> for the server-side of the rpc

23:39 <TMM> push requests on a thread pool