Update, December 5th: a deployment was made to the grid. Server update version 2019-12-04T20:29:26.533447 comprises:
- A build of release 2019-11-15T21:13:13.532828 using new build technology.
- Addresses some cases of scripts erroneously stopping.
- Fixes a crash.
Hopefully, this addresses the recent issues with scripts failing until a region is restarted (see BUG-227864).
Please refer to the server deployment thread for news and updates:
- On Tuesday, 3rd December, the planned Main SLS channel deployment was cancelled. Had it been deployed, it would have comprised server maintenance release 2019-11-19T22:26:38.532992, built with the updated server build tools, and which includes:
- Improved crash detection during shut-down.
- Fixes for some race conditions with LSL scripts that could cause them to handle the same event more than once during a roll.
- Updates to improve simulator security.
- The planned RC deployment for Wednesday, December 4th, comprising server update has also been cancelled.
The reasons for the cancellation are related to the return of recent issues with scripts failing until a region is restarted (see BUG-227864) and which has been an intermittent issue recently, with Simon Linden noting:
We originally had planned to update the main channel regions with the version that’s on the RC channels, but held off as we are looking into issues about scripts not running. It appears to be an old issue that’s always been there but is worth investigating … we’re trying to get an update together for tomorrow morning [Wednesday, December 4th]. It was a very small percentage [of scripts affected] but that’s enough to be painful when there are millions running.
– Simon Linden
Simulator Performance Issues
There is a forum thread concerning simulator overload issues that occurred over the Thanksgiving weekend (see: Huge intermittent sim overload). These issues were further confirmed at the SUG meeting by others seeing similar problems on other regions. Linden Lab do not currently have an explanation for the problems, but Maxidox Linden, one of the Second Life QA team members, did take a moment to address claims in the thread that LL “not caring” about issues like this:
I’m going to let the devs here handle this topic but I just want to chime in that I’ve read that thread and we absolutely do care about issues like this, and don’t put our development time solely into new feature work (I’m the one who tests a bunch of the stability and performance change.
– Mazidox Linden
The discussion on this topic raised questions on the simulator statistic Simulation Time and what it actually records, given it is often tied to simulator issues – as is the case with this issue. Exactly what it records has been a subject of discussion for around the last decade, and came up again in terms of whether it might point to a possible cause. After taking a look, Simon Linden noted:
So I did some digging. “Simulation Time” seems to be a catch-all category that includes odds and ends like dealing with terrain layers, some physics synchronization, parcel info. So not a lot to use tracking this issue down.
There have been no official viewer updates to mark the start of the week.