FYI quitter.no is now on "nightly" branch. I have also implemented restart of daemons and servers at 00.00 every night. There are 8-12 threads handling the queue to coop with the traffic. (approx 1000-2000 notices pr.hour - varies through the week). Working on getting RabbitMQ as the queue handler, however I receive notices and it is pushed into the RMQ queue, but it's not pulled back from the queue into the GS database. Looking into it again when time. !quitterno
quitter.no finally back again. Sorry for the outage. the local hackerspace have increased the mem on the instance, hopefully it will be more stable going forward. !quitterno #quitter !gnusocial
FYI: Now quitter.no should be stable again after cleaning up 14mill rows. I’m just impressed the whole system survived this. No reboots, restart services etc. Thanks to @mmn for help today! !quitterno !gnusocial
FYI: a bit slow on quitter today – the queue of notices has a lot to do now, caused by deleting a user fired off a lot of notices to subscribed users. (your subscribed users noticed will appear after a while…) !quitterno
#observation : data portability has gone from bad to worse as far as !GNUSocial is concerned. Previously it used to issue some sort of a "backup" file even though I failed to use it in any meaningful way such as importing to another instance. Now it doesn't work at all. I get no file when I click on the backup link here on !quitterno and also on #gnusocialno I know it has never been a priority in GNUSocial. But right now we got more and more users usingpublic sites. Making user data portable it important. Even #Twitter offers this, even though you cannot migrate that data to any other server, you can open and view your #Twitter updates locally in a browser.