The open source metaverse

Update 2015-01-05

With the holiday breaks behind us, we expect the data center to receive the recovered drive from the recovery service soon, but do not have an ETA yet. We hope to have an ETA or the drive itself before the next update.

Both new asset servers were fully configured and brought online temporarily working as a master-master replication pair.
This was the first full-scale test of the new database replication, robust xassetservice/hgassetservicece and nginx all together, and it was more successful than we’d hoped.
The database, asset server, and nginx all started together without error, and successfully replicated assets from master to slave in both directions.
This means that asset transactions will be replicated in real-time to the standby asset server, regardless of which server we’re using to serve assets.
The next tests involve loading a test copy of Wright Plaza, testing from master to replica and back to master in both directions, checking all OSgrid operators are happy with failover, failback, and that everyone is satisfied the new asset server design will be able to handle the flood once the gates are opened.
Some technical hurdles remain to be worked out and we’re consulting with core OpenSim developers as needed to remove these as OSgrid comes online using an entire new asset service.

OSgrid restart is getting closer but there is still no specific date or ETA as we carefully test this entirely new asset service.

Restructuring discussions have begun with an eye to rebalancing the plaza loadout, backups, and the various OSgrid services. At this stage its more inventorying the existing resources and services, and scoping out what to leave along and what should change for 2015 and beyond.

Next Update by 2015-01-11 or 2015-01-12.