Server Deployments Week #47
As always, please refer to the server deployment thread for the latest updates.
- There was no deployment to the Main (SLS) channel on Tuesday, November 17th
- The maintenance package deployed to all three RC channels in week #46 received a further update on Wednesday, November 18th. This was apparently due to the postponed deployment from week #46 not containing all of the desired updates.
After mysteriously vanishing from the viewer release channel, the HTTP project viewer (Project Azumarill) reappeared on Tuesday, November 17th, still with the same version number (22.214.171.1246796, dated November 2nd) as previously.
The Vivox viewer entered the release channel as a RC viewer on Tuesday, November 17th. Version 126.96.36.199.8.7.307744 corrects a number of Voice quality and connection issues on both Windows and the Mac.
The Chrome Embedded Framework Viewer also reached RC status with the release of version 188.8.131.527894 on Wednesday, November 18th.
Group Ejections and Group Chat
SVC-32, “Person kicked remains in chat session” came up for discussion at the Simulator User Group Meeting on Tuesday, November 17th.
This is the ongoing issue where those who are ejected from a group whilst in a group chat session still being able to continue using the chat until they close the window (with various permutations on the issue vis muting and unmuting people).
The cause of the problem seems to be quite deeply rooted in how the group chat service works, with Simon Linden noting, “I hate that one … and I’ve been lurking in some large groups (like FS) and seeing it abused. It’s an ugly problem and I’ve looked into it … basically because the info about how to add the logic isn’t where the log needs to be run … I looked into revoking caps and other ideas but didn’t find a silver bullet, or a pile of them, that would work … I’ll try to get some time to see if there’s anything that can be done without major architecture changes or new servers.”
Object and Linkset Contents and Limits
Changes were recently made to how many objects can be placed within an individual in-world / worn object and linkset. The limit per individual objects is now 10,000, and the limit per linkset is 30,000. The limits were revised to help deal with people using objects which massive contents from eating up server resources and impacting region performance as a form of griefing. However, a side issue with the changes has been that tools such as texture organisers with huge contents spread through the linkset can stop working – permanently.
To help people check their organisers and make suitable adjustments in line with the new limits, the Unpackistan region has a 90K limit for linksets (although there was an issue with this region as well – BUG-10695 – which should now have been corrected).
Concerns were raised the simulator User Group over the lack of follow-through when limits like this are changed (there has been no Wiki update in this instance, for example).
While it is understood limits may be changed to help reduce problems with griefing, etc., and thus the Lab may no wish to announce forthcoming changes in case it leads to an increase in abuse before the door is shut, so to speak, it was pointed out that providing some indicators as to what has changed, and ensuring reference documentation is properly updated could go a long way to help people figure out what might have happened as a result of such changes having an unexpected repercussion (as in this case).
Acknowledging this point, Simon agreed to take the concerns raised back to the office for further discussion.