I also agree with Alex when it comes to ssh .. That's still a big issue .. load balancing would be nice too ..
regards Werner On Thu, Mar 27, 2014 at 9:33 AM, Alex Mandel <[email protected]> wrote: > Eggs in one basket? > > Why not load balance 2 servers in different parts of the world? > Bandwidth at OSUOSL is unlimited, hardware is already paid for (It's > raid SAS drives, not cheap hardware). > > Sharing people time also helps. I was the only one awake when this issue > came up because I'm in the western US. I'd also rather see SAC and QGIS > admins work together, not just divide into territories. > > I got involved with SAC as my way to contribute to QGIS originally, but > now find myself guessing and without the ability to help on the new > servers at all. So I'm all for experimenting with the right balance, and > hope we can collaborate. Part of the point of the shared hosting is that > QGIS server admins benefit all OSGeo projects if there's an incentive to > work together. > > I'd prioritize upgrading Redmine and fixing the missing ssh key > management over moving any other sites. That's the site that cause the > most trouble in terms of hosting. Then there's things like CDN/DOS > protection, SPDY optimization, https for all downloads, caching. All > sorts of better ways to spend time than just shuffling static > html/sphinx all over the place. > > Thanks, > Alex > > > > On 03/27/2014 12:11 AM, Paolo Cavallini wrote: >> Why not moving also the website to the new server? >> All the best. >> >> On 27 marzo 2014 08:03:35 CET, Richard Duivenvoorde <[email protected]> >> wrote: >>> On 27-03-14 03:45, Alex Mandel wrote: >>>> So Nathan noticed the website wasn't coming up. Specifically >>>> "The requested URL /en/site/ was not found on this server." >>>> >>>> I believe the issue is disk space. I'm guessing the sphinx is built >>>> elsewhere and transferred in on a schedule without checking how much >>>> space is left on the server. I got to this conclusion by noticing >>> that >>>> other languages are working fine. >>> >>> Yep disk was full again. Site is still being build on the same server, >>> and ends with a move/relink/remove-old step (apparently even when the >>> build process died in the middle)... Have to fix that. >>> >>> Fact that only english was affected was because only english and polish >>> (new) was being rebuild... >>> >>>> Looking at the space: >>>> There are 10G of backups in various places, all current >>>> There's also 6G of apache logs of various sorts. We should find >>>> somewhere to archive things over a few months old. >>> >>> Thanks for that. Untill now we moved plugins.qgis.org to new server, >>> but >>> are waiting for some time to remove old one. >>> I'm also trying to clean up now. >>> >>>> Temporarily I'm moving all *.gz files in /var/log/apache2 over to the >>>> osgeo downloads machine in my user home folder. Frees up 1.5 GB, >>> leaves >>>> last 2 logs for all services intact (ie log and log.1) >>> >>> Thanks again, we need a plan for this. >>> >>>> I did not touch the www_qgis.org folder inside apache2. >>> >>> English is up again now. >>> >>> Thanks. >>> >>> Regards, >>> >>> Richard >>> _______________________________________________ >>> Qgis-developer mailing list >>> [email protected] >>> http://lists.osgeo.org/mailman/listinfo/qgis-developer >> >> >> >> _______________________________________________ >> Qgis-developer mailing list >> [email protected] >> http://lists.osgeo.org/mailman/listinfo/qgis-developer >> > > _______________________________________________ > Qgis-developer mailing list > [email protected] > http://lists.osgeo.org/mailman/listinfo/qgis-developer _______________________________________________ Qgis-developer mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-developer
