Disabled by default. Currently can only be enabled with console "debug lludp oqre start" command, though this can be started and stopped whilst simulator is running.
When a connection requires packet queue refill processing (used to populate queues with entity updates, entity prop updates and image queue updates), this is done via Threadpool requests.
However, with a very high number of connections (e.g. 100 root + 300 child) a very large number of simultaneous requests may be causing performance issues.
This commit adds an experimental engine for processing these requests from a queue with a persistent thread instead.
Unlike inbound processing, there are no network requests in this processing that might hold the thread up for a long time.
Early implementation - currently only one thread which may (or may not) get overloaded with requests. Added for testing purposes.
Also closes behaviours on disconnect instead of interrupt, though this makes no practical difference.
If existing behaviour is None, other added behavious will not take affect until None is removed (as this is an infinite wait until interrupted).
If n > 1 for RootTerseUpdatePeriod only every n terse update is actually sent to observers on same region, unless velocity is effectively zero (to stop av drift).
If n > 1 for ChildTerseUpdatePeriod only every n terse update is sent to observers in other regions, unless velocity is effectively zero.
Defaults are same as before (all packets are sent).
Tradeoff is reduction of UDP traffic vs fidelity of observed av mvmt.
Increasing n > 1 leads to jerky observed mvmt immediateley for root, though not on child, where experimentally have gone to n = 4 before jerkiness is noticeable.
Rapid polls are more expensive than triggered events (several polls vs one trigger) and may be problematic on heavily loaded simulators where many threads are vying for processor time.
A triggered event is also slightly quicker as there is no maximum 200ms wait between polls.
This is already going to be correctly set by WaitForUpdateAgent() earlier on in that method, which is always called where a callback to the originating region is required.
This kind of polling is very expensive with many bots/polling threads and appears to be the primary cause of bot falloff from the client end at higher loads.
Where inbound packet threads can't run in time due to contention and simulator disconnect timeout occurs.
Experimentally, on the Linden Lab grid the avatar can rotate slightly before triggering AvatarUpdates, whereas this is practically impossible in OpenSimulator.
These updates allow other avatars to see rotations, though sensitivity is low since other avatars can only be seen in one of 8 body rotations.
This commit changes sensitivity from 0.01 to 0.1, which better matches LL and reduces UDP traffic which has a beneficial impact on network and CPU load.
This has no impact on rotations in the simulator itself so simulation fidelity is the same as before.
To change this setting back for test/other purposes, edit RootRotationUpdateTolerance in the [InterestManagement] section of OpenSim.ini
These govern when AgentUpdates are sent to observers on position, rotation and velocity changes to an avatar (including the avatar themselves).
Higher values reduce AgentUpdate traffic but at a certain level will degrade smoothness of avatar and perceived avatar movement.
Allows experiments in manually reducing updates under heavy load.
Activated by "debug scene set client-upd-per" console command.
In a simple test, can send as few as every 4th update before observed movement starts becoming disturbingly rubber-banded.
Corresponds to ResendAppearnceUpdates setting in [Appearance] in OpenSim.ini
This was originally implemented to alleviate cloud appearance problems but could be too expensive with large numbers of avatars.
This governs when child agent position changes are sent to neighbouring regions.
Corresponding config parameter is ChildReprioritizationDistance in [InterestManagement] in OpenSim.ini
For test purposes.
This covers event queue setup messages and some outgoing messages (e.g. EnableSimulator)
In my experience these messages are only useful if you really know what they mean and you're looking for them
Otherwise, they're quite spammy.
Event queue DebugLevel 1 is enabled with the "debug eq 1" console command
On non-HG this is on the only recognized failure state so we can return more information in the error result.
On HG there are multiple failure states which would require more work to distinguish, so currently return the unsatisfying "Internal Error" like some other existing calls.
This gives a count of all requests made to the remote inventory service.
This is finer grained than inventory.httpfetch.ProcessedFetchInventoryRequests since such a request can be comprised of many individual inv service calls.
In addition, this will count requests that don't go through the HTTP inventory fetch (e.g. HG, archiving, etc.)
Conflicts:
OpenSim/Services/Connectors/Inventory/XInventoryServicesConnector.cs
A stop gap solution - a better one may be to improve stats display on simulator-side.
Caps information is still accessible via the "show caps stats by user" and "show caps stats by cap" commands
For each service endpoint (e.g. posts to the xinventory service), a stat is available which shows the number of requests received and moving average per second
The full name is "service.<http-method>:<path>.requests (e.g. service.POST:/xinventory.requests)
I forgot that a null 'their version' would not be passed over the wire and ends up as an empty string instead (like older simulators).
So instead pass through the correct simulator protcol version instead (SIMULATOR/0.3) when querying from login or hg login.
Also removes a debug console write for agent limit accidentally left in for the same commit.
Relates to mantis 7276
Despite the comments in the code, it appears that the issue where the .NET performance counter was wrongly idle time time on Mono was fixed in 2009.
https://bugzilla.novell.com/show_bug.cgi?id=468625
Which means that the workaround is no longer necessary and produces bad results instead.
Justin Clark-Casey (justincc)