Server

Pathfinding

Pathfinding is closer to release. I expect we will see it arrive on the Release Channels this month.

Andrew Linden says they feel that are now feature complete for the initial release. I suppose this means that all the features the Lab believes are important for successful use of Pathfinding have been added. They still have bogs and issues to sort out. But, the feature set is fixed until after initial release.

They have decided to change the Pathfinding Tools in the viewer. Seems it is a work flow change that drives the need for viewer changes.

If you have used the viewer to edit Navmesh, you realize there are some klunks. Unfreezing the mesh, editing, and Freezing the Navmesh have some delay issues that have to be handled. Once a Navmesh has been changed and locked it takes some time for the server to recalculate the Navmesh. As Andrew says, may be tens of seconds. With multiple people editing the same Navmesh in a region you might imagine what problems a 20 to 40 second delay might cause them.

Read more

There is a bit more news on the Thursday service outage. A firmware upgrade in the routers is the likely problem. Other data centers have seen similar problems. Today, Monday, they downed the system from about 7 to 9 AM PDT. The revised upgrades presumably made it into firmware and we are back up. And that is about all we are ever likely to know about the event.

Magnum Release Channel

This channel is running the update package many Second Life® users are waiting for. This is the one with PRIM_SLICE, HTTP header size control, and llGetAgentList(). Unfortunately this package is having problems and is not likely to make it to the main grid.

The problem seen is attachments fail to attach/rez. In some cases they rez but the scripts fail to load.

Read more

Seems we had a major melt down Thursday morning (USA morning). One of the Linden Lab® data centers had hardware network failures. In systems this size there is usually considerable effort put into load balancing to distribute the workload. So, with the center dropping offline everything went out of balance as the servers remaining online could not carry the load. The Lindens turned off some services until they could get the data center back online. Mostly money related services went offline. It took several hours to correct the problem.

During that time login’s were closed or inaccessible and the main Second Life® web sites went offline.

There is no truth to the rumor that one of the gerbils peed on a router after waking up from a hard night of partying in Zindra.

Main Channel

This week a roll did make it to the main grid. This is the package with the Inventory API’s. So, with luck we will see better inventory and delivery performance. As Thursday the Tuesday roll has not shown any problems.

Read more

This week a new package rolled out to the main grid. This is the package with the Inventory API changes. These are intended to make inventory transfers more stable and reliable. Features also will allow the Inventory loads to scale better. This is the package that was running on Blue Steel and Le Tigre release channels.

The main grid also got the change in flight height. The limit has been moved from 150-200m to 5,000m. Flight feathers should no longer be needed. However, some feathers will make you fly faster and that may be a worthwhile reason to use one.

If you have a flight feather or whatever device that sends you into orbit, contact Simon Linden

Release Channels

The package running in Magnum failed this past week. Group notices being sent from regions running Magnum failed. This package will continue to run in Magnum with a number of additional fixes.

Read more