Second Life project updates 36/1: server, viewer

Le Avaline Village; Inara Pey, August 2015, on FlickrLe Avaline Village August 2015 (Flickr) – blog post

Server Deployments Week 36

As always, please refer to the server deployment thread for the latest information / updates.

  • On Tuesday, September, 1st, the Main (SLS) channel received the server maintenance package deployed to all three RC channels in week #35, comprising:
    • A fix for BUG-9504 “Clicking on any object that affects the navmesh while in Mouselook dirties the navmesh”
    • Internal simulator fixes
  •  On Wednesday, September 3rd, the BlueSteel RC received an updated version of the server maintenance package first deployed (and subsequently rolled back) in week #34, which comprises internal fixes aimed at improving inventory performance.

Due to the issues experienced when this latter package was deployed to all three RC channels (such as the “zombie eyes” situation), the package is only being deployed to the one RC; Magnum and LeTigre will remain unchanged from week #34, keeping them on the same release as the Main channel.

SL Viewer

On Tuesday, September 1st, the Mesh Importer RC viewer updated to version, making its promotion to the de facto release viewer in week #36 unlikely, but not impossible.

Region Restarts and Caps Failures

A problem often encountered following region restarts is that some regions come back with a caps failure (so a lot of things that should work, don’t). While less frequent an occurrence than has previously been the case, the problem does still occur. The problem is thought to be at the server level, as regions hitting the problem tend to all be located on the same server.

Commenting on the matter at the simulator User Group meeting on Tuesday, September 1st, Simon Linden said:

I have a good theory about caps failure on the rolls but the last time I tried to fix it, the update went badly and we rolled back :). My theory is good, the side effect was bad.   When we restart regions, we do them all at once.   My fix was to pace that slightly, and not overwhelm the caps system.   However, the delays confused the system starting the grid, and it started the same regions multiple times, which didn’t go well. And of course it didn’t do that on the beta grid.

Since his initial attempt at correcting things, Simon has been engaged on other work (such as getting group chat fixed), but he is hoping to get back to working on this problem at some point in the future.