Server Deployments week 51
Tuesday 18th December saw the Main Release channel get the maint-server release as deployed to Magnum in week 50, as Maestro Linden indicated might be the case during the Server Beta meeting on the 13th December. There have been some reports of teleport issues following the deployment, but these are thought to be a network-side issue, rather than a problem with the server code itself.
All three RC channels should be receiving the same package on Wednesday 19th December, 2012 which is also the Magnum deployment from week 50, with a single additional bug fix.
Release notes (for Magnum, but applicable to all three RC channels)
Forum thread for the deployments.
This is the last set of server-side deployments for 2012.
SL Viewer Updates
The release version of the official SL viewer rolled to the 3.4.3 code on Tuesday 18th December, with the release of 126.96.36.1998262. Among a range of updates, this release include the major GPU table updates seen in recent beta viewer releases, the additional graphics settings and the tiling fix for high-resolution snapshot (MAINT-638). This release also sees the return of JIRA numbers to the release notes, matching the recent beta viewer release notes update.
The beta viewer should be moving to the 3.4.4 code with the next release, although it is not clear whether this will be before or after the holiday period.
Interest List Update
Andrew Linden has been tweaking and optimising his Interest list code. The first phase of the new interest list code is aimed at reducing the amount of information being sent to the viewer by the server. Speaking at the Simulator User Group meeting on the 18th December, Andrew commented, “Last week I managed to find an optimisation opportunity in my Interest List project, which shaves a few hundred usec off of some cases where there are a lot of Full Object Updates. So this would translate to better server FPS when in large groups of avatars or when in the presence of stuff that is generating lots of full updates.”
As I reported last time, object updates take three forms, Full, Terse and Delete. Full updates are generated when an object is seen for the first time, or which is undergoing a constant change in appearance / position. Terse updates are sent only when an object changes its appearance / position. Under the current Interest list system, both types of update are sent to the viewer, regardless as to where the viewer actually needs to process them. A primary part of Andrew’s work has been reworking the code so that updates – particularly terse updates – are only sent to the viewer when the object generating them is in the camera’s field-of view. This new optimisation work should further help reduce the number of updates being sent to the viewer regardless as to whether or not it has to process them.
Avatar / Server-side Baking
Since the announcement that the this project is now entering what is effectively a viewer testing phase, some developers have been carrying tests with the updated server-side code on Aditi, together with the Sunshine project viewer. Some issues have been noted in obtaining updates when changing outfits, leading to some off artefacts appearing in the viewer (available from the Sunshine wiki page).
This problem is actually one of a number of issues already identified within the system, and while there have been suggestions as to how the issue might be fixed, the Lab appears confident that they have a grasp of the problems and that their architecture and approach is such that they’ll be able to correct the problems in what is still very much an initial alpha testing phase (with Nyx referring to the viewer-side code as “pre-alpha”.
There is still no real news on this project, only that it is “coming together” and that there is likely to be news “soon”. Whether that is pretty soonTM or real soonTM is open to debate – but either way, it is unlikely much will be forthcoming until early 2013.
Work on this from the Lab’s side of things appears to be stalled, for reasons which aren’t clear.When asked at the Open Source Dev meeting on Monday 17th December if there had been any progress on developing a test plan, Oz Linden replied, “”Not really, I’m afraid. But I have not forgotten about it.” He then went on to add, “We’ve got some folks that will be focusing more on avatars from a few different perspectives after the New Year, and I think this will be part of that.”
Precisely what this means is unclear, and Oz refused to be further drawn. While there are a number of open JIRA relating to the avatar, the context of his statement would tend to suggest whatever the “different perspectives” may be, they are related directly to the deformer, and so are likely to be linked to things like avatar weighting, etc., rather than to anything particularly fundamental with the avatar itself.
In the meantime, and while a further request for test clothes hasn’t openly been made recently, it appears the Lab have only received “some” samples since the last call was made. Therefore, if you do wish to contribute clothing, please refer to Oz’s original request on the subject. Any samples which are offered should be uploaded using the very latest version of the deformer project viewer.
Advanced Creation Tools Permissions
July saw the launch of the first phase of the Advanced Creation Tools, also referred to as experience tools. Following problems with an initial deployment of the tools in June, which resulted them being exploited as a means of griefing, the “first phase” of the release saw the tools implemented with existing permissions system in place, with the intention of updating the permissions system to allow the tools to be more fully used “in the future”.
Since that initial deployment, there has been little news on the project or what is happening, permissions-wise. However, speaking at the last Simulator User Group meeting of 2012, Simon Linden confirmed that while the work had been “stalled” for a time, it has again, “Been getting attention recently.”
Region Performance Issues
Toysoldier Thor raised a forum thread on this issue a while ago and I’ve reported on it a few times in these weekly reports. Essentially, events attracting large numbers of users experience a sudden onset of a range of symptoms, including network performance drop-outs, increased lag and people experiencing region disconnects.
Similar issues were noted at the Content and Simulator User Group meetings during week 50. However, whether these were down to the same issue is potentially questionable. Both of the User Group meetings were subject to a protest by a group over a situation (now apparently resolved), some of whom were noted as carrying abnormally high script loads; one individual within the protesting group alone was counted with 648 active scripts totalling 23,664KB and grabbing 15.792 ms of frame time – and issues within the meeting started shortly after their arrival. So it is possible that the sheer script load was as much responsible for issues during the User Group meetings as anything else.
It is also interesting – though possibly coincidental – to note that the Phoenix / Firestorm Office Hour meeting on the 15th December had between 70 and 80 avatars in the region, which had scripts disabled at the estate level, and no performance issues of the kind experienced elsewhere were reported during the meeting.
In the meantime, and while no single cause of the problem has been located, the code deployments for week 51 includes some fixes – one directed at “low-level threading issues” which might help stabilise the situation.
Issues around logging-in to Aditi were raised at the Simulator User Group meeting, and it was again confirmed that where the account management database capacity issues are concerned, the solution is still on the acquisition of additional hardware (most likely from within LL) in order to increase the database capacity.This is still going to take time to complete.
In the meantime, the separate issue of general account synching between Agni and Aditi continues, with users frequently having to make numerous password changes to their account in order to get the two services to correctly synch. It is unclear as to what the underpinning issue here might be, but Andrew Linden commented during the Simulator User Group meeting that another LL dev was working to correct what Andrew referred to in quotes as “memory leaks” within the Aditi database which may be linked to the Agni / Aditi synching / import issue(s).
For now, the only solution to getting accounts to synch between the two grids is to repeatedly change your password (and allow between 24 and 48 hours between attempts). How this will affect the proposed Threaded Region Crossing pile-on tests has yet to be seen. If you are intending on helping with that test, and have not logged-in to Aditi recently, you may want to check whether you can, and if not, reset your password now, in order to allow the systems to hopefully synch.