[Opensim-dev] Observations on XMLRPC timeouts

[Kurt Taylor]

This is also very similar to what I have been seeing on Linux and Windows.

Re: test results - while I agree that ad-hoc testing shows a completely new
set of problems, I think a lot can be said for having a base set of tests
with specified configurations that can be run across all environments.


Kurt R Taylor

(SL: Kurt Stringer)
http://opensimulator.org
International Business Machines, Corp.




                                                                           
             "Nebadon Izumi"                                               
             <nebadon2025 at gmai                                             
             l.com>                                                     To 
             Sent by:                  opensim-dev at lists.berlios.de        
             opensim-dev-bounc                                          cc 
             es at lists.berlios.                                             
             de                                                    Subject 
                                       Re: [Opensim-dev] Observations on   
                                       XMLRPC timeouts                     
             06/13/2008 09:32                                              
             AM                                                            
                                                                           
                                                                           
             Please respond to                                             
             opensim-dev at lists                                             
                .berlios.de                                                
                                                                           
                                                                           




My results are virtually  identical to what Melanie says above, my testing
has all occured on OSGrid.  One thing i can tell you is these results are
very difficult to reproduce under controlled situation, one thing that is
key to really bring these results to the forefront is chaos, an
uncontrolled totally open environment such as OSgrid is really suffering
when linux is in the mix.  Windows on .NET seems unphased by most things
thrown at it.   I can tell you it should not matter  what version of linux
or what version of mono you run, results are consistent almost across the
board.    I only suggest we throw caution to the wind with controlled lab
environment lab results, as many who run in very tightly controlled
environments dont seem to see exactly the same issues.

Nebadon

On Fri, Jun 13, 2008 at 3:47 AM, Mariusz Nowostawski <
mariusz at nowostawski.org> wrote:
  Melanie wrote:
  > I have tested OpenSim in 5 different combinations today, in the same
  > grid, same UGAIM.
  >
  > This is a grid that is affected by a number of issues related to
  > grid comms.
  >
  > Tested:
  >
  > linux - mono
  > Observed numerous failures of MapBlockQuery and other XMLRPC calls.
  > Also, inflated memmory footprint.
  >
  > windows - cygwin - mono
  > Application did not run. Regions did not load, communications did
  > not work
  >
  > windows - native - mono
  > Observed numerous failures of MapBlockQuery and other XMLRPC calls.
  >
  > windows - native - .NET
  > No observed issues.
  >
  > windows - cygwin - .NET
  > No observed issues
  >
  >  From this, i conclude that the XMLRPC issue is a mono issue, while
  > the memory bloat is a Linux issue

  Hi Melanie,

  We are running similar tests with setups on Windows, Solaris and Linux
  in our lab. To be able to compare apples with apples, would you be able
  to provide more details of what exact configuration of your linux/mono
  and windows/.net configs were? In particular, what exact mono version
  have you used. with what exact garbage collector, what was the linux
  kernel, and how was the exact memory usage between different setups?

  I think it would be valuable to establish the same common UGAIM setup
  with a single common region, that different people would be able to use
  for testing (a sort of like an agreed benchmark). There are many groups
  looking into opensim performance and reliability but we have no way of
  comparing, replicating or even re-evaluating the results, because many
  factors contribute to a particular server behaviour, memory usage and
  performance.
  If you could use one of the publicly available regions that we could use
  too, than we could cross-validate our results, and try to pinpoint
  exactly where the problems are, and how we can fix them. Our group would
  be more than happy to help investigating where the problems are exactly.

  --
  best regards
  Mariusz
  _______________________________________________
  Opensim-dev mailing list
  Opensim-dev at lists.berlios.de
  https://lists.berlios.de/mailman/listinfo/opensim-dev
_______________________________________________
Opensim-dev mailing list
Opensim-dev at lists.berlios.de
https://lists.berlios.de/mailman/listinfo/opensim-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20080613/117f45ef/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20080613/117f45ef/attachment-0003.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pic13203.gif
Type: image/gif
Size: 1255 bytes
Desc: not available
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20080613/117f45ef/attachment-0004.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20080613/117f45ef/attachment-0005.gif>