MantisBT - opensim
View Issue Details
0008185opensim[GRID] Hypergridpublic2017-06-07 02:212017-06-09 12:08
aiaustin 
 
normalminorsometimes
newopen 
PCWindows10
master (dev code) 
master (dev code) 
opensim-0.9.0-1465-ga23f993
Grid (Multiple Regions per Sim)
BulletSim
.NET / Windows64
None
N/A
0008185: [LLUDPSERVER]: UseCircuitCode handling... Exception Object reference not set to an instance of an object
Incoming Hypergrid visitor ... red error showed, though avatar appeared to arrive okay...


[LLUDPSERVER]: UseCircuitCode handling from endpoint 90.200.161.115:57878, client <firstname.lastname> @hg.osgrid.org <UUID> failed. Exception Object reference not set to an instance of an object. at OpenSim.Region.ClientStack.LindenUDP.LLUDPServer.HandleUseCircuitCode(Object o) in d:\Temp\opensim-0.9.0-1465-ga23f993\OpenSim\Region\ClientStack\Linden\UDP\LLUDPServer.cs:line 1742
No tags attached.
Issue History
2017-06-07 02:21aiaustinNew Issue
2017-06-07 10:34UbitUmarovNote Added: 0031998
2017-06-07 10:54aiaustinNote Added: 0031999
2017-06-08 02:40UbitUmarovNote Added: 0032000
2017-06-09 12:08aiaustinNote Added: 0032023

Notes
(0031998)
UbitUmarov   
2017-06-07 10:34   
this happens all the time?
(0031999)
aiaustin   
2017-06-07 10:54   
I have just seen this today @Ubit, once... but it is unusual to see red errors on my grid consoles... so I thought I should report it.
(0032000)
UbitUmarov   
2017-06-08 02:40   
Hard to tell, from normal code paths that should not happen (as mantis 8183)
unless the avatar was on the region before and something went wrong when it left
our handling of those problems is still pretty bad.

Several red messages can now show up on shutdown, since as needed to work around mono bugs more threads are explicitly killed and modules disposed.
Most code has no shutdown concept.
Friends for example will go on trying to do 1000 http connections to your 1000 friends..
(0032023)
aiaustin   
2017-06-09 12:08   
Just to note, the red error was as a Hypergrid visitor arrived in the grid... it was running normally... not in a shutdown stage. Could have been a one off glitch. Will watch out for other instances.