00:01 <samba_> ok, thanks. isd I think the backup feature is not really useful to export/import a mediawiki db

00:02 <samba_> is just in the case I need to restore the grain, but I've not the xml I need to import the source

00:02 <samba_> and I've no easy access to the db

00:02 <samba_> so, for what I can see the only option seems to do an Specia:Export/Pagename for everypage

00:05 <ocdtrekkie> I'm wondering if there's any sort of global category you could use in the top box of that Export page to grab everything.

00:05 <ocdtrekkie> But generally, yeah, that seems to be the case. :/

00:06 <isd> The other option is to pick open the zip file and try to run that php script on the contents from somewhere, though it might take a bit of fiddling

00:06 <ocdtrekkie> Export all pages is a common sense thing to do, and I'm a bit mystified why MediaWiki chose not to offer that in their web interface. They've gotten better over time, but MediaWiki software seems to have a preference for telling you to do things in the file system on the server for a lot of tasks.

00:07 <isd> but there's going to be a snapshot of the db in there somewhere.

00:08 <ocdtrekkie> Might not even be that hard. It looks like it might be sqlite.

10:55 <Laxmin> Hello!!

12:13 <samba_> helo

12:13 <samba_> I need to restore a pad

12:13 <samba_> I've the backup but I'm not sure where is physically the file

12:14 <samba_> I've something like : sandstorm/var/sandstorm/apps/0059cac5e958a9bd57b41f24d7b56503

12:14 <samba_> and I don't know which id is pointing to the pad I want to restore

12:14 <samba_> how can I know that ?

12:38 <TimMc> samba_: This is a filesystem backup, not a grain backup, right? (Or... is there a difference? I've never checked.)

12:42 <samba_> a fs backup TimMc , I need to check if I can restore a specific pad

12:47 <TimMc> So you have a grain you want to revert back to that version?

12:49 <TimMc> In just a bit I'll have my sandstorm box back up and running and will be able to give you more details, but there's a directory of grains down in the sandstorm directory, and worst case you can find your grain by putting a unique string in the etherpad and grepping. :-P

12:49 <samba_> so, there's no way to know which grain is app 1010219292901219819218 ?

12:50 <samba_> where sandstorm store this data?

12:54 <TimMc> samba_: Ah, here we go. Do you know the grain ID?

12:55 <samba_> no, that's the problem

12:55 <samba_> I've the url or the name of the pad

12:55 <TimMc> hmm

12:56 <TimMc> FYI the grains are stored in /opt/sandstorm/var/sandstorm/grains/ as directories, by grain ID.

12:56 <samba_> the grainID is the same text I see in the url bar when I open a grain?

12:57 <samba_> because if so, I found my pad to restore in : sandstorm/var/sandstorm/grains/AQ8TzrCHScFvJAQN9K4mYX

12:57 <TimMc> yes

12:57 <TimMc> And there should be a ./sandbox/sqlite3.db in there if it's etherpad

13:02 <samba_> there is

13:03 <TimMc> OK! Sounds like you have a path forward now. :-)

13:04 <samba_> yes, I'm doing it..

13:04 <samba_> all this stuff I'm doing is because there was an issue with etherpad+sandstorm I was not able to recreate and debug

13:05 <samba_> a pad (very big) was edited adding some info last week, then yesterday checking this info some user told me the modification was gone

13:05 <samba_> so, I tried to do the same , but the pad seems working

13:05 <samba_> now I'm thinking about to do a restore and check it out

13:05 <samba_> do you know if there's any issue with big pad and sandstorm?

13:14 <TimMc> Nothing I know of.

14:12 <ocdtr_web> Added an app porting bounty for Monica on http://sandsheep.com/bb though I might try it myself.

14:13 <ocdtr_web> Couple people on the Reddit discussion were interested in the idea of it being on Sandstorm.

18:09 <isd> TimMc: re: the fs vs. grain backup distinction: a grain backup is just a zip of the grain directory (so includes logs + the sandbox dir, which is the grain's /var)

19:32 <TimMc> hah

23:39 <isd> So I'm trying to figure out how to correctly implement postStreaming/putStreaming in my Go wrapper for web-session. Is it actually possible (without promises, which the go capnp implementation still can't return afaik) to implement a handler for those that expects to read in data from the request body before returing any of the response? The RequestStream docs suggest that the caller will always call getResponse before actually

23:39 <isd> g any data. So I don't see, without promises, a way to do this.

23:39 <isd> Which is a problem.