Otter Lake, June 2019 – blog post
Server Deployments
- There was no deployment to the SLS (Main) channel on Tuesday, August 13th.
- On Wednesday, August 14th the RC channel should be updated as follows:
- BlueSteel and LeTigre should be updated to server maintenance update 19#19.08.07.529856, containing internal updates which should see improvements to the percentage of scripts running per CPU cycle (particularly for Full regions) and reduce the portion of a cycle given over to idle scripts.
- Magnum should be updated to server maintenance update 19#19.08.06.529800, containing internal fixes.
SL Viewer
The Love Me Render viewer, version 6.2.3.527758, formerly the Rainbow RC viewer dated June 5th, was promoted to de facto release status at the start of the week.
The rest of the official viewer pipelines remain as follows at the time of writing:
- Release channel cohorts (please see my notes on manually installing RC viewer versions if you wish to install any release candidate(s) yourself):
- Umeshu Maintenance RC viewer, version 6.2.4.529755, August 7th.
- EEP RC viewer, version 6.4.0.529247, July 12th.
- Bakes on Mesh RC viewer, version 6.3.0.529185, July 11th.
- Project viewers:
- 360 Snapshot project viewer, version 6.2.4.529111, July 16th.
- Legacy Profiles viewer, version 6.2.3.527749, June 5. Covers the re-integration of Viewer Profiles.
- Linux Spur viewer, version 5.0.9.329906, promoted to release status 29th November 2017 – offered pending a Linux version of the Alex Ivy viewer code.
- Obsolete platform viewer, version 3.7.28.300847, May 8th, 2015 – provided for users on Windows XP and OS X versions below 10.7.
In Brief
Scripts
- Script usage is still seen as possible cause of contention (e.g. because one parcel owner is consuming all script resources within a region). A means has been suggested for changing (see BUG-225391), but the Lab see the issue as complex and so wanted to get improvements such as those being initially deployed to the BueSteel and LeTigre RC channels issued first before potentially digging more deeply into the question of script usage and allocation.
- How clear a picture does the Full private region Top Scripts give for script usage in a region? Obviously, it can help, but as Mazidox Linden noted: “I do something similar all the time. It doesn’t give you the fine grained details you’d imagine for our set-up however. For that I use other (mostly internal) tools/techniques.”
- It’s been suggested that script information could be extended to parcel level. This is something the Lab has indicated it is musing doing – but no firm plans to implement any changes if the Lab go in that direction have been indicated. It is also a complicated issue, as Simon Linden summarised:
So how would you treat someone with 1/4 of a region and a club [and] 50 avatars come to see a popular performer, and the other 3/4 are empty at the time? Give [the club] all the script time? Or leave 75% for the … other land?
Region Memory Bloat
As a reminder: a region can suffer significant memory bloat during repeated terraforming. Simply put, this is the result of the navmesh (part of the region’s physics) being reformed during the terraforming. This can have numerous impacts: overall performance can be affected, as can rezzing objects. Nominally, if a Full region’s memory allocation hits approx 900 Mb, or a homestead 230 Mb, it will shut down rezzing (if the underpinning simulator gets within ~90% of its memory limit (256MB for a Homestead & 1024MB for a full region) it will disallow rezzing).
So, if you are running a region that starts reporting issues with rezzing and have been terraforming, consider a restart (or requesting a restart from support).