There is a bit of news coming from various places.
Server Side Appearance has a problem that the Lindens have been working on for the past week. Well, thinking about, I’m not sure they are writing code. I first mentioned the problem of corrupting modifiable eyes and skin when going in and out of SSA enabled and disabled regions here: SSA Coming Not Too Soon.
Some people say the problem is unique to the Phoenix Viewer. Latif Khalifa tried the SL Viewer 1.23, pre-SSA Singularity, Imprudence and Phoenix. But, only found the problem on Phoenix. Whirly Fizzle said. She could reproduce the problem on the SL Viewer 1.23. Whirly says old FS viewers were OK and did not have the problem.
Nyx Linden made a special trip to the Sever Beta meeting to discuss the problem, saying: “Thanks to the support from the Phoenix/Firestorm team, we’ve been able to identify the cause of that issue. We’re looking into what options are available to us.”
That leads me to think they were still thinking the problem through on Thursday (6/6).
When asked if the bug would hold up release of a Release Channel version Nyx said, “We’re still evaluating what impact it will have and what options there are to mitigate that. We haven’t announced a timeframe for taking SSA to a release channel.”
In a more private meeting word is out to third party viewer developers that regions in AGNI do have SSA enabled. So, it isn’t holding up release as much as it is holding up announcement of it. TPV Dev’s are testing in both AGNI and ADITI. This allows TPV QA teams to more easily test. But, making that information public would probably create some chaos.
The new LSL functions llReturnObjectsByOwner and llReturnObjectsByID are running in ADITI. Look for the CCMTEST11 channel. That meas they are getting close to moving into a release channel on AGNI. See: New LSL Functions Coming.
You already know there was no roll to the main channel this week. I did see some region restarts in the main channel. I don’t have any information on how extensive that was or wasn’t. I have a monitor in my SL home that records restarts on my land.
I’ve already explained what went to these channels in: Second Life News 2013-23. That did not change.
What is new information is the priority of the changes in Blue Steel and Le Tigre. The code to fix the massive disconnections from region servers is running in all three release channels. The BS-LT regions have the current server software with the Disconnect fix. Magnum is running the server software with the Interest List and Disconnect fixes.
The desire is to get the Disconnect fix rolled out, hopefully with the Interest List fixes. By running a version without the Interest List they increase the chance that the Disconnect fix will make it out next week.
If you have a region in Blue Steel or Le Tigre and you are still seeing disconnects, get your information to the Lindens via a JIRA and mention Maestro Lindens name. Also mention JIRA BUG-2564 and the forum thread Increase in Instant SIM LAG & Crashes During Larger Events – Network Source?
JIRA BUG-2564 – Users are randomly disconnected from Frodes sim.
Description: Multiple users are being disconnected from Frodes sim at frequent random intervals. This can affect from one to many or even all users in Teqi’s Lounge. On my viewer the screen turns grayscale and I get a message saying “You have been disconnected from Second Life…” On restart I can sometimes log immediately back into my last location. Other times I get a message saying the server is still logging out my account.
This is happening more frequently. On 5/11/13 it happen more than 3 times between 10:00pm and 11:00pm SLT. On 5/12/13 it happen 4 times between 8:00am and 8:30am SLT.
This is a new issue and is being reported to me by many visitors to Teqi’s Lounge on Frodes.
If this problem is persisting then the Lindens did not get the fix and they need to know. If they don’t hear any complaints, they will rightfully assume they have it fixed.
Coming next week will be a fix for bandwidth throttling. A symptom of the problem is LSL scripts that use HTTP-In randomly lose their URLs. It has also been confirmed that if the bandwidth isn’t throttled properly (which is what you see with this bug), it would result in packet loss, which could result in the viewer sending more resend requests. That would consume more bandwidth, which could result in more packet loss, which would result in more resend requests… a death spiral.
This bandwidth throttle problem also affects the downloading of large scripts when a user attempts to edit them. Large in this case being near the script’s text limit of 64k. Rumor says this can be worked around by lowering the bandwidth setting.
The fix is currently on ADITI in the DRTSIM-208 channel. Regions in that channel are Gibson and Noyo.
Do you see invisible avatars? Right. Seek professional help… no… you know what mean. There are a few reports of this problem. If you are detecting the problem and can reproduce it, file a JIRA.
Some people see this happen when they use SL Viewer 1.23 and come down from a high elevation.
A failed TP may also derender all avatars. Others see a render fail after TP, even if it succeeds.