The Beta Server meeting this Thursday was definitely attended by more Lindens than I’ve ever seen in any one place.
- Baker Linden
- Caleb Linden
- Dan Linden
- Don Linden
- Kelly Linden
- Kurt Linden
- Log Linden
- Maestro Linden
- Nyx Linden
- Simon Linden
That is a pretty big list of Lindens.
Maestro covered the server rollouts for the week. No surprises there. But we did get a bit of information on a couple of the fixes in the Magnum RC channel.
From the previous news article you know that the main grid got the Multi-Threaded Rez. This lets the server do other things while lots of objects are being downloaded for rez in the viewer. Big link sets were a problem. So were avatars with lots of attachments. Maestro tells us that this change will help any region where lots of objects are being rezzed, including avatars with attachments.
However, most of the improvement will be for the rezzing of mesh objects.
For Magnum the big fix is the ‘User not online’ bogus error message. It seems this was introduced with a fix for VWR-786 – IM does not respect “allow user to see my online status” flag in friends list.
This is feature request described as:
If I have someone in my friends list who I would rather avoid for a while, I can *untick* the box that allows that person to see my online status. However, if that person IMs me then they can instantly tell that I’m online, as they don’t get the standard offline IM response.
I would have liked to add this as “major”, as it makes it socially impossible to use the online status hiding functionality. However I thought I’d get shouted at if I did 🙂
I assume that the fix would be pretty easy — just check to see if the person being IMed is hiding their online status from the person sending the IM, and if they are then show the “user is not online” text.
So, this update (786) is being backed out. The Lindens will rethink it and try again later.
Over in ADITI, the preview grid, there are new server packages running. The Interest List changes running in Solariam (Adult) and Tischeriidae (Moderate) are:
- Fix an interest list issue where you ‘lose track’ of a vehicle after region crossing.
- Fix an interest list issue where off-camera objects would suddenly ‘snap’ into place when you turned your camera. Off-camera object updates will now be sent at a low rate (instead of not at all).
- Fix an interest list issue with ‘sim teleporters’, where an object that got moved by a ‘teleporter’ object would still appear at the original position to other avatars.
- Some improvements for object loading.
These fixes have passed the QA tests and will likely move to the RC channels next week.
Monty told us about HTTP changes coming down the road. These changes are in place in some regions of the ADITI grid. So, we can test how these changes are working. Monty tells us the first three test channels with regions for DRTSIM-203 are up. The DRTSIM-203 channel is what is targeted for RC in AGNI.
The three versions of the HTTP changes are DRTSIM-203: Normal, Happy, and Angry. They are described this way:
- DRTSIM-203. Normal release intended to go to Agni supporting keepalive connections and other changes. Regions:
- TextureTest2. High texture count, no meshes, low residency limit to prevent interference when doing benchmark testing.
- (Coming soon) MeshTest2. High mesh count (many distinct mesh assets which all look the same), few textures. Low residency limit to prevent interference when doing benchmark testing.
- DRTSIM-203H. Our ‘Happy’ channel with more generous limits and optimistic service controls.
- TextureTest2H. Identical to TextureTest2.
- (Coming soon) MeshTest2H. Identical to MeshTest2
- DRTSIM-203A. Our ‘Angry’ channel with stricter and preemptive enforcement of limits (generates many 503 errors).
- TextureTest2A. Identical to TextureTest2.
- (Coming soon) MeshTest2A. Identical to MeshTest2
Test regions share object and texture references so if you are trying to measure download rates or really exercise the network, you’ll need to defeat caching. Typically with a restart and manual cache clear or your own preferred method.
Aditi is also hosting some of the server-side baking work and you may not get a reliable avatar appearance unless you use a Sunshine project viewer.
What the Lindens are looking for on these channels is:
DRTSIM-203. HTTP issues generally. HTTP texture download reliability and throughput. Script writers using llRequestURL and llRequestSecureURL should try to get an A/B comparison going between a normal ‘Second Life Server’ region on Aditi and DRTSIM-203. Particularly with competing traffic like large texture or mesh downloads. Scripts aren’t getting a boost with this work but they shouldn’t be adversely impacted. Mesh also isn’t getting a boost this time but, again, negative impact should be avoided. Third-party viewer developers should test for overall compatibility with all HTTP services.
We’re interested in reports of regressions in any areas. We *are* expecting more 503 errors (0x01f70001) in log files as it will be possible to push requests faster than before and certain throttles will be hit. As long as these are recoverable, they’re not a regression, they’re an indicator of better utilization.
These are not sandbox areas, so no building. This is partially to keep the regions uniform. This allows comparative tests between the Normal, Happy, and Angry regions.
Also the regions have limited residency. Meaning only about 3 people can get into a region. That prevents avatar loads from interfering with measurements of the HTTP performance.
Viewer developers will be using the regions to test their viewers. The Lindens will be using them to collect performance data. I don’t know that average users will have much use for the regions, so we should mostly stay out.
However, if they are empty and you want to see how they work, go for it. Do use the SSAB Project viewer. I understand it has the latest HTTP code. Nyx says there is a new version fo the SSAB Project viewer coming out soon, like Friday.
The HTTP protocol is becoming more important to SL. Textures can download via HTTP or UDP. If the HTTP fails for a texture the viewer and server will work to fall back to UDP. But, there is no fall back to UDP for mesh objects. HTTP failure will affect prims, sculpties, and meshes. So, this is a server that REALLY needs to work well.
Non-mesh and non-texture code that uses HTTP will get a boost from these changes. The changes are about using keep-alive-connection which are about getting the next 100 or 1000 things. In the future pipelining will be part of the SL HTTP protocol.
IPv6 is not going to be part of the change for now. Too many consumer routers do not support IPv6. So, that is being put off.
Zero Script AO
Maestro is telling us that next week Kelly will tell us how to have a zero script AO… I have no idea if that comment was an inside jole, real, or something else. But, if it is possible I would like to know how.
Pile-On Test of SSAB
Nyx Linden and crew ran the pile-on test in the sunshineSLS regions. It seems like they were happy with all the data they were getting.
Things went pretty well with the exception of attachments. Attachments started to fail to attach. Lots of Failed to Rez error messages were popping up.
We will likely hear more about the test at Nyx’s Monday meeting.