This reverts commit 59c0df962a.
This is necessary to preserve the casing of column names, otherwise pgsql converts everything to lowercase.
TIL, Unlike mysql, a quoted column still refers to the column itself...
This may well be the cause behind "operator does not exist: uuid = character varying" pgsql errors.
If so, the other instances will also be changed.
Relates to http://opensimulator.org/mantis/view.php?id=7398
This is primarily to fix an issue where inventory items for NPC attachments would not be saved in archives as the assets referenced in the appearance notecard were not checked in this manner.
Relates to http://opensimulator.org/mantis/view.php?id=7376
This UUID gatherer provides a superset of the previous gatherer's functionality
as it also allows the caller to control gathering iterations for load purposes.
closer to SL. This change should address small floor edges acting like
walls, approaching a step at any angle (other than walking backwards)
will allow walking up, and reducing the avatar pop-up when going up
stairs.
Note that I also made things consistent by removing both the EstateDataService and the SimulationService into their own dlls, just like all other services. They really didn't belong in Services.Connectors, since everything in that component is about network connectors to robust backends. We may have too many dlls, and at some point it might not be a bad idea to merge all services into one single dll, since they all have more or less the same dependencies.
Change how physics engine is created in unit tests to resolve a lib reference problem.
Add ShapeInfoInfo class to collect info about the created physical shape
for debugging and unit test testing.
them in the code for later use rather than just reverting them.
Fixed the throttle tests for the new algorithm used when packets
are marked as expired.
- It establishes 4 digits for opensim versions
- It uses the same number between opensim releases and mono addins versions
It also eliminates the last addin.xml files that were still there, for consistency.
than drop exponentially to 0 (and then adjust up for the minimum flow), drop on
the delta between current rate and the minimum rate. This should smooth the fallback
to minimum.
This cleans up Opensim's use of mono addins. In particular, the extension points /OpenSim/RegionModules and /OpenSim/WindModule moved from OpenSim.exe to OpenSim.Region.Framework.dll. From here on, developers of region modules should declare their dlls to be dependent on OpenSim.Region.Framework, starting with version 0.8.1
Additional changes:
- Addins version uniformly updated to 0.8.1. These numbers should be compatible with the release numbers or else it becomes very confusing.
- Mono addins directives moved from files addins.xml to embedded directives in the class and assembly declarations, to make it all consistent
adaptive throttle by a full MTU. This is consistent with some implementations
of congestion control algorithms and certainly has the effect of opening
the throttle window more quickly after errors. This is especially important
after initial scene load when the number and size of packets is small.
throttles. Setting adaptive_throttle_min_bps will change the
minimum rate that the adapative throttles will drop to in case
of network packet loss. The current rate default rate is 256kbps.
The viewer can throttle rates under that amount, but the dynamic
adaptation will not.
algorithm for dropping packets is a modified two state algorithm for creating
bursts of dropped packets. As configured there is about a 1.5% drop rate.
Invocation of the packet loss code is commented out by default.
Normally, WhiteList is an empty list but from Mantis 7389 it looks like it might sometimes be possible for it to be null (haven't seen this up till now)
It looks like failing with an exception instead of properly replying to the request (which comes in via a cap) might be enough, surprisingly, to freeze a viewer until timeout.
Part of http://opensimulator.org/mantis/view.php?id=7389 but probably unrelated to the actual issue of that mantis.
This was beause the code was finding the script DLL compiled for the source region as everything is in the same appdomain and using this as the location for the destination script state, etc.
This resolves the regression by passing the proper destination separately from the DLL retrieved.
Probably a regression since commit d7b92604 (11 July 2014).
Added regression test for this case.
At least partly addresses http://opensimulator.org/mantis/view.php?id=7278
If log level is 1 then every script load is logged.
This means the <logger name="OpenSim.Region.ScriptEngine.XEngine"> section in OpenSim.exe.config is no longer needed to avoid log spam on regions with many scripts and can be removed.
Setting m_CurrentCompile = null in the finally block reduces the risk that an exception could permanently stop any future scripts compiling until the simulator is restarted.
If an exception is seen from this then please report and further changes to fix the bug or improve compiling reliability can be made.
This commit addresses the following issues were causing velocity to be set to 0 on the new region, disrupting flight in particular
* Full avatar updates contained no velocity information, which does appear to have some effect in testing.
* BulletSim was always setting the velocity to 0 for the new BSCharacter. Now, physics engines take a velocity parameter when setting up characters so we can avoid this.
This patch applies to both Bullet and ODE.
Instead, the failure information is effectively added to the memory cache (so no persistence over simualtor sessions).
A future improvement may be to invalidate negative cache results after some time has passed in case the failure was transient.
Looks to resolve http://opensimulator.org/mantis/view.php?id=7382
This has not been maintained for more than 2 years, is unimplemented for newer features (e.g. built-in groups) and has no core developers using it.
If somebody fixes these issues then it could be reinstated.
This is to achieve a clean separation of concerns - the watchdog is an inappropriate place for work management.
Also adds a WorkManager.RunInThreadPool() class which feeds through to Util.FireAndForget.
Also switches around the name and obj arguments to the new RunInThread() and RunJob() methods so that the callback obj comes after the callback as seen in the SDK and elsewhere
This is implemented with a new collision type (PhantomToOthersAvatar) to potentially allow colliding and non-colliding avatars to be present in the same scene.
So there is no provision yet for giving avatars different collision types.
This commit replaces the temporary change in commit f3eaa6d8 where avatars would never collide when using BulletSim
This is equivalent to the av_av_collisions_off option in ODE.
This improves upon the earlier naive simply queueing immplementation.
Threshold is 30 seconds. If this happens to a user they can relog and fetch will be reattempted.
This is to enable an imminent change where incoming HG scene object fetching can assess the time taken by each request rather than being forced to perform all requests in one call.
Soon, this will replace the existing UuidGatherer since it is both simpler and more flexible.
This is to reduce the potential for overload of the threadpool if there are many simultaneous requets in high concurrency situations.
Currently only applied to AvatarProperties and GenericMessage requests.
This shows named threadpool calls (excluding timer and network calls) that are currently queued or running.
Also shows total of labelled and any anonymous calls.
Instead of processing all incoming attachment scene object concurrently, process them consecutively to eliminate potential overload from this source.
This is a naive implementation because it does not currently account for slow foreign asset services.
Although it may take longer, this approach may also improve attachment visibility for HG avatars
since the scene object is now always added to the scene after receiving assets from the foreign service and not before.
This is to avoid user confusion in the oscc rehearsal as they are often not aware that this fails because no e-mail is set.
Also may be failing in the hypergrid case, though this may also be a config issue.
This is meant as a temporary solution.
At least on Mono 3.2.8 (but not under Windows), one can bind multiple UDP sockets to the same port by default.
Different simulators cannot demultiplex each other's messages, so a set of confusing non-obvious errors arise if this occurs.
This change prevents such multiple binding.
This shows summary wearables information (shape, hair, etc.) for all avatars in the scene or specific information about a given avatar's wearables.
Similar to the existing "attachments show" command.
In "show throttles", also renames 'total' column to 'actual' to reflect that it is not necessarily the throttles requested for/by the client.
Also fills out 'target' in non-adapative mode to the actual throttle requested for/by the client.
This is done by adjusting the order of code so that SceneAgent will always be set before adding the client.
Various parts of the code (rightly) assume that a a client registered to the manager will always have a SceneAgent set no matter what.
When an HG avatar enters a scene, it delays processing of entity updates. Could be crowding out by other updates or something else.
This delay in ones own av mvmt updates results in mvmt lag experienced on the client. Avoiding the internal LLClientView for these packets appears to resolve this issue.
Appears most noticeably for avatars with attachments, though has also been seen on those without sometimes. Hasn't been observed for non-HG avatars in general.
Will be investigating exactly what the problem is, at which point there will be a more permanent solution.
Information is also available in "show server throttles" but that's more for non-debug info rather than attempting to get and set parameters on the fly for debug purposes.
This is because objects with lots of parts can have a lot of xml to load into memory, and this has been seen to have a noticeable performance impact.
Whereas streaming has been seen to reduce the impact in normal serialization.
Implmentation is messy but I couldn't see a better way of doing it when you can't assume that you know the exact structure of the input XML.
To do this required GetMesh to be converted to a BaseStreamHandler
Unlike GetTexture connector, no redirect URL functionality yet (this wasn't present in the first place).
This has two aims
1) Reduce initial teleport failures when a foreign Hypergrid user enters a region by not holding up the teleport for attachment rez (this can be particularly costly when HG gets all assets in the object graph.
2) Reduce server load that may impact other simulator activities.
This complements existing JobEngine options that perform initial login attachment rez and appearance send in consecutive tasks.
This was because specifying a max client throttle would always request the max from the parent server throttle, no matter the actual total requests on the client throttle.
This would lead to a lower server multiplier than expected.
This change also adds a 'target' column to the "show throttles" output that shows the target rate (as set by client) if adaptive throttles is active.
This commit also re-adds the functionality lost in recent 5c1a1458 to set a max client throttle when adaptive is active.
This commit also adds TestClientThrottlePerClientAndRegionLimited and TestClientThrottleAdaptiveNoLimit regression tests
This only had one child, which is the 'adaptive' token bucket.
So from testing and currently analysis, we can use that bucket directly which simplifies the code.
This is separate from the user-oriented "show throttles" command since one will often only want to know about varying client throttle settings.
Currently displays max scene throttle and adaptive throttles config if set.
This is the total of queued outgoing packets across all connections, as also seen in the "show queues" command.
Gives some early indication of whether the simulator can't send all outgoing packets fast enough.
Though then one would want to check that this isn't due to a few bad client connections.
In theory, there should be no difference between these mechanisms.
However, on at least Mono 3.2.8 waiting via an event appears to be much more accurate.
Job engine is controlled via "debug jobengine start|stop|status".
Can only currently be enabled and disabled dynamically at runtime.
Currently only applies to code sending initial region data (objects, other avatar data) to a client.
This involves making PhysicsActor.PIDActive get as well as set.
On physics components that don't implement this (all characters and some phys engines) we return false.
Same behaviour as on Linden Lab grid.
Will probably also address http://opensimulator.org/mantis/view.php?id=7369 by not generating spurious object updates when llStopMoveToTarget() is called in attachments where the avatar is not moving.
Unfortunately, it's not currently easy to do this with "max-agent-limit"
- this must be separately set as MaxAgents in region config if it's to persist over restarts.
This current allows one to set two region parameters
agent-limit <int> will set the current root agent limit for the region, as also settable through the viewer, though some impose a max setting (e.g. 100).
max-agent-limit <int> will set the maximum allowed root agent limit. This can also be set via the MaxAgent parameter in region config.
This is somewhat more in keeping with something like osForceAttachToOtherAvatarFromInventory()
and potentially allows a separate osForceSit() command with High threat rather than VeryHigh that only sits the owner and can be enabled without enabling sit of other avatars.
Allows a script IN the target prim to force an avatar to sit on it using normal methods as if called by the client.
Overload method of osForceSit() to allow a script NOT in the target prim to force an avatar to sit on the target prim using normal methods as if called by the client.
This patch is based on previous work from
http://opensimulator.org/mantis/view.php?id=4492
and also includes the suggestions from justincc including change of threat level
Thank you Christos Lightling.
This code originates from when IEntityTransferModule was shared rather than one per region.
Now it's one per region we know that callers are always in the same scene as the module.
This is cinderblocks' transaction_result.diff from http://opensimulator.org/mantis/view.php?id=7329 but I have used lsl.parser.cs and lsl.lexer.cs files generated directly from opensim-libs rather than those supplied in the patch.
I also added scriptEvents.transaction_reuslt.
The required parser/lexer generation file changes were made in commit d564f28 in the opensim-libs repo.
Thanks!
Attachment persistence is not handled in this way and this just results in a load of busy work until a check in each SOG terminates a backup check for attachments anyway.
This setting was originally added some time ago to deal with issues where appearance was not received properly by all users.
However, it does not scale well with large numbers of agents.
Disabling to see if the original problem has abated or whether this will have to be tackled in another way.
[Startup] default setting UpdateTempCleaningEveryNFrames becomes UpdateTempCleaningEveryNSeconds.
Default becomes 180s instead of effective 182s (which would also vary with any changes in frame time or extra long frames)
Unlike "debug lludp packet" which logs at the point where OpenSim first asks the clientstack to send a certain outgoing packet, this logs immediately before the actual send.
For low-level debugging purposes.
This allows one to log the packets received by a particular bot that are not duplicates of already received packets.
Similar to the OpenSimulator command at the same name but currently any positive level logs all received packets.
No facility yet for logging outgoing packets.
For debug purposes.
The functionality is the same but this allow us to monitor such tasks via "show threads" and abort them for test purposes, etc.
Also extends thread names to provide more info (e.g. SendInitialDataToClient says what client the task is for).
Other parts of OpenSimulator are relying on SP.Velocity == 0 for vehicles.
So add and use SP.GetWorldVelocity() instead when we need vehicle velocity, along the same lines as existing SP.GetWorldRotation()
As per http://wiki.secondlife.com/wiki/LlSetStatus
Setting STATUS_BLOCK_GRAB_OBJECT prevents or allows move of a physical linkset by grab on any prim.
Setting STATUS_BLOCK_GRAB prevents or allows move of a physical linkset by grab on a particular prim.
Previously, setting STATUS_BLOCK_GRAB would prevent drag via all prims of the linkset.
This was previously effectively being done by XmlDocument in the multiple passes through the XML.
This change tells XmlReader to ignore whitespace. This also means changing arguments to use XmlReader instead of XmlTextReader (a descendent of XmlReader) directly.
XmlReader.Create() has been the recommend way to create XML readers since .NET 2.0 as per MS SDK and is the only way to specific ignore whitespace settings.
Reading large XML documents (e.g. complex attachments) is CPU expensive - this must be done as few times as possible (preferably just once).
Reading these documents into XmlDocument is also more resource intensive than using XmlTextReader, as per Microsoft's own publication "Improve .NET Application Performance and Scalability"
Optimization of other cases will follow if this change is successful.
This is to see if an inaccuracy in sleep times under load is responsible for increase in frame times even when there is spare time still available.
Can currently only be activated by setting "debug scene set update-on-timer true".
Can be switched between timer and thread with sleep updates whilst the scene is running.
However, many locks remain since they may effectively be providing transactionality in some operations (e.g. prim updates across multiple tables).
These are candidates for being replaced with proper database transactions, since this would not block unrelated operations (e.g. land save and object save)
or unrelated operations on the same tables (e.g. storage of one linkset whilst another is being removed).
In practice, any performance deg due to contention is probably rare and short lived as the major prim operations are performed in memory and only persisted some time afterwards.
These locks are not necessary since the connection is taken from the underlying mysql pool and not shared.
Such locking is already not done by some other parts of OpenSim.Data.MySQL.
Pointed out by arribasim-dev
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.
This adds the "show stats", "stats record", etc. commands and information on available Threadpool threads, etc.
It also adds the Watchdog which logs warnings if time between executions is unexpectedly large.
This is to avoid issues where many bots connect to a single end point with multiple regions, where each region requires a long-lived poll connection for each bot.
This lock serialized all requests and made the inventory throttling in WebFetch redundant.
By moving this lock, two simultaneous requests may now take place which may help with http://opensimulator.org/mantis/view.php?id=7054
Move the experimental extra features functionality into the GridService. This sends default values for map, search and destination guide, plus ExportSupported control to the region on startup. Please watch http://opensimulator.org/wiki/SimulatorFeatures_Extras for changes and documentation.
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.
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 may have been the trigger CheckSendingPatchesToClients() dictionary out of sync exceptions in today's load test.
Don't need to check ContainsKey() since Remove() returns false on a request to remove a key that it doesn't have
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 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.)
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
This is for testing purposes (chiefly http://opensimulator.org/mantis/view.php?id=7054) so many not be permanent
Setting this will change the xinventory request timeout from the simulator from the default 100 seconds
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)
Fixes http://opensimulator.org/mantis/view.php?id=7280
It can't be done this way because the stats data needs to show up on the console at all log levels, not just debug.
But this means setting it to log at fatal, which is not appropriate for this stuff in the log.
I understand the desire but this has to be done some other way, perhaps by (yet another) config parameter.
Also, this was already being done with the ClientStatsReport but that also should be done in another way, I think.
This reverts commit 5d53412766.
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.
Another manifestation of this bug is that after a Hypergrid teleport, when you click on one of the avatar's attachments the object doesn't show its name. This means that the viewer knows the attachment is there, but the simulator does not.
The problem was caused by treating Hypergrid teleports as if they're Logins (because the teleport flag ViaLogin is enabled).
This may fix: http://opensimulator.org/mantis/view.php?id=7238
This may fix: http://opensimulator.org/mantis/view.php?id=7220
This caused the client to stop responding, and even the simulators to have problems. The solution is to disallow crossing before the previous cross has completed.
The "Sculpt" field in prims is used for both Sculpties (where the assets are Textures), and real meshes. Meshes require a different download URL than textures.
When sending an ImprovedInstantMessage to a group, the IM's binary bucket is supposed to contain the group's name (this is what SL does). Singularity uses this to show the group name when it shows the message at the bottom of the viewer for a few seconds: "[Group Name] From User: Message". Before this update, the group name was empty ("[]").
This update doesn't have any visible effect in Firestorm, because it doesn't use the group name sent in the IM.
This, by default, enables terrain patches being sent to each avatar
from the avatar away (rather than the old outside-in pattern), only
sending terrain patches within the avatars view distance (making
view loading quicker), and sending multiple terrain patches per
protocol packet (making terrain loading and editing quicker).
when registering a new region.
Adds parameter "[GridService]SuppressVarRegionOverlapCheckOnRegistration=false"
that can be turned on to suppress the error check if a simulator's database
has old regions that overlap.
co-op should be more stable as it doesn't abort threads, which can trigger virtual machine instability
This change will be invisible to users as script DLLs are recompiled automatically where necessary, though the change won't take affect until the next simulator restart.
This change has no effect on existing script state.
If you want to continue using abort, set ScriptStopStrategy = abort in the [XEngine] section of OpenSim.ini
As per http://msdn.microsoft.com/en-us/library/system.random%28v=vs.100%29.aspx, the .NET Random class is not thread-safe.
If called by multiple threads at once, methods may return 0.
Except for llRand(), other OpenSimulator code did not lock before calling a shared Random instance.
This commit adds a ThreadSafeRandom class that extends Random but does internal locking so that it is thread-safe.
This change is invisible to existing callers and the explicit locking in the llFrand() implementation is now redundant.
We have to do this since we can't unload existing DLLs if they're all in the same AppDomain.
But we can still update the underlying DLL which will be used in the next simulator session.
Code was identical apart from error logging, but if there are failures creating these directories then you'll be
seeing lots of errors anyway, and these will be more informative
In commit e6080a38 (Wed Mar 19 00:29:36 2014) I renamed this from "debug stats record start|stop"
Unfortunately, I didn't do this fully so before this commit "stats record start|stop" will report a usage failure with the old debug text.
Unfortunately this is in the 0.8 release. The workaround is to repeat the last command twice (e.g. "stats record start start")
Justin Clark-Casey (justincc)