Yang, circa 10/4/08
I have been asked to take up the weekly update for OpenSim on this blog. I will start by describing my test setup for OpenSim on OSGrid which has been running for over a year. This blog will not describe the plazas, but rather the next set of test regions next to Wright Plaza on the SouthWest diagonal corner. For this purpose, I personally have six sims running on three different servers. I will be referring to these by name in the weeks to come and here is the test setup.
Yang – A single region running on a 1.6Ghz AMD CPU with 2GByte of RAM on a Linux Fedora desktop on the floor of my office at home.
Io, Europa, Phobos & Deimos – Four regions running on a dual-core Intel 3.4Ghz CPU with 3Gbyte RAM running on a different Linux Fedora desktop also on the floor of my office at home. These are collectively referred to as the moons.
Celt – A 256MByte VPS with a single core Intel Xeon running at 2.5Ghz and no swap from bodhost.com in Durham, UK.
Yang and the moons, five regions total communicate across my ADSL in Irvine, CA. Celt is seperate and in Europe. These six regions are on the SouthWest corner of Wright Plaza on OSGrid and form a 2×3 rectangle. Yang has about 2000 prims and 50 scripts and is the most heavily developed. Phobos is next, then Celt and the other three moons are mostly empty at this point.
Revisions r6543, r6550, r6564, r6620, r6669, r6672 & r6675 have been recently tested. There have been half a dozen crashes on the six regions in the last week, mostly due to thread stuff. OGP is currently suffering a bit with recent changes including sim crossings.
We have changed our official client for testing to 1.20.16 as that is the new download on secondlife.com and since updating, adjacent regions have not been visible. With the latest r6675, now tagged as release 0.5.11, I can now see and cross into adjacent regions again from Yang. R6675, aka 0.5.11 seems stable and I would recommend it.
Script implementation and testing continues with about 56 LSL functions left to implement and patches are solicited. Additionally, many have been able to run kan-ed scripts with no problems. Google “grobe kan-ed lsl” for the kan-ed scripts.
I would recommend some attention be given to implementing some of the stubs of packet handlers at the bottom of LLClientView.cs. I believe we are at the point where we can benefit from additional packet handlers.
Sincerely, Charles Krinke, Irvine, CA, Core Developer and Director of OSGrid