[VideoPlayer] Failure accessing d3d device

[viiiper]

GRAW SERVER RUNNING 1.35 DEDICATED.

Anyone had this error or can GRIN explain it

it happens every time a particular custom map is played and the cycle moves to

another map. The crash log on the dedicated server says the map that is being loaded is

the problem but it's a stock GRIN map ie: tdm01, tdm09 etc....

THe map plays both sides and then the next map is loaded (or tries to load)

Crash in application version: p6 - 19 oct 17.42

[VideoPlayer] Failure accessing d3d device

SCRIPT STACK

data\lib\setups\setup.dsf(0)

data\levels\tdm05\tdm05.dsf(0)

===========

Crash in application version: p6 - 19 oct 17.42

[VideoPlayer] Failure accessing d3d device

SCRIPT STACK

data\lib\setups\setup.dsf(0)

data\levels\tdm09\tdm09.dsf(0)

Could it be that the video card is out of memory ? or something else.. GRIN ??

I'm asking a question... thx

[FI_FlimFlam]

We are starting to see the same thing with our dedicated server with custom maps. 2 in particular confrontation and tirania aldea. I find this strange since supposedly the dedicated server shoudln't need d3d api's. If anyone can provide any informtion on this it would be helpful. This started happening around the March 8th and has been the main reason for server crashes since.

Crash in application version: p6 - 19 oct 17.42

[VideoPlayer] Failure accessing d3d device

SCRIPT STACK

data\lib\setups\setup.dsf(0)

data\levels\custom_levels\tirania aldea\tirania aldea.dsf(9)

[FI_FlimFlam]

I think I'm narrowing it down. after deleting the one map and then adding a new custom one, the server still crashes. But it crashes on the same slot in the rotation rather than the same map. So, it's crashing on the 5th map every time. I am testing the server with a reduced number of maps to see if it's related to the number of maps loaded or perhaps it's related to total size of maps loaded. I'm not sure. I'll post back as I find out more.

[viiiper]

Do not fear Flimflam we are working on the problem at the BDA, your welcome on TS to chat.

BDA 2007 SERVER

no pwd

66.29.71.168:8962

We have eliminated many things and are hopefully close to a solution, but we know one

thing it's not the slot position. It is defiantly tied to certain maps and it crashes the server

with an error log referring to the next map in rotation (being loaded) but the problem

is always the map that was played previously.

It also seems to be tied into version 1.35 (editor) and how is makes maps or the game

engine it self on how it deals with maps. Grin in their wisdom as usual make changes and

do not inform us (there customers) of all the new features that worked in previous

versions and now do not work.

I will update this thread when we get to the bottom of GRINS modifications.

viiiper

john

p.s. interesting fact is the error log when written too shows the error on line (9) of the next map

Crash in application version: p6 - 19 oct 17.42

[VideoPlayer] Failure accessing d3d device

SCRIPT STACK

data\lib\setups\setup.dsf(0)

data\levels\custom_levels\test135\test135.dsf(9)

Which when you look at the dsf file involved is:

1)import Application, World, Underlay, Global;

2)import Math, Sound;

3)import setups.Setup;

4)import utils.dev.editor.WorldHolder;

5)

6)object _world;

7)

8)init() {

9)Setup.start_loading_video(); <-----------

We have used 3 diffrent dedicated standalone servers to isolate this problem and

it seems to be with 1.35, we fresh installed graw on one and patched to 1.35 with

the same results.

[viiiper]

OK after over 30 Hrs of multi testing of decompiled map files

I have found our problem and maybe yours:

The mentioned fault strangely originals from the file ZONES.XML

which can be found in an unbundles map in

....data\levels\custom_levels\map_name\xml

The offensive line in this file is

<borders tl="x y z" br="x y z"/>

Where x y z are values that mark out the boarder that defines the region that will

become the mini map. It seems that this line was required in earlier versions of the

GRAW EDITOR.

I hope this helps you FLIMFLAM and anyone else that experiences this error log

crash listing

SOLUTION: remove/remake this line from the file ZONES.XML

The maps we were testing with the problem were all TDM, but this does not mean

that the problem is limited to TDM only.

example of working zones.xml

<zones>

<zone name="alpha" base="1">

<center pos="0 0 80000"/>

<rect corner1="-22400 -15000 0" corner2="-18100 -16800 0"/>

</zone>

<zone name="bravo" base="2">

<rect corner1="-9400 -800 0" corner2="-6800 -3400 0"/>

</zone>

</zones>

example of none working zones.xml

<zones>

<borders tl="-26000 -1000 5000" br="-6700 -17300 5000"/>

<zone name="alpha" base="1">

<rect corner1="-22400 -15000 0" corner2="-18000 -16900 0"/>

</zone>

<zone name="bravo" base="2">

<rect corner1="-9400 -800 0" corner2="-6800 -3400 0"/>

</zone>

</zones>

alternate working zones.xml

<zones>

<borders tl="-26000 -1000 5000" br="-17300 -6700 5000"/>

<center pos="0 0 80000"/>

<zone name="alpha" base="1">

<center pos="0 0 80000"/>

<rect corner1="-22400 -15000 0" corner2="-18100 -16800 0"/>

</zone>

<zone name="bravo" base="2">

<rect corner1="-9400 -800 0" corner2="-6800 -3400 0"/>

</zone>

</zones>

REMEMBER the values quoted in the above examples are unique to one map, it's the general layout

we are talking about.

best regards

viiiper

john.

p.s. if there is an alternate solution from GRIN, I'd like to here it... Are they limits to the Boarders statement in terms of value or sign + / - (can all four X,Y values be negative) ? Does the "center pos" statment in ZONES.XML

mean anything in TDM etc....

[FI_FlimFlam]

That makes complete sense.

But let me throw a wrench in it for a second.

Since I have noticed the crash at loading of the map in slot 5, the server has been regularly crashing at about 3 hour intervals. I had been removing what I thought were offending maps. I haven't attempted to remove the map in slot 4. The last thing I did this afternoon was remove the last 3rd of the map rotation from the list. (about 9 or 10 maps) while preserving the same first 2/3rd's. Guess what, no crash. Last I checked it was on map 7 or 8 of the rotation. Perhaps this is indicative of 2 separate issues, but by chance have you tried reducing the rotation in half or so to see if it stabilizes? I'm going to do some more testing tomorrow and see what is going on and see if I can reproduce it reliably.

Edited March 11, 2007 by FI_FlimFlam

[viiiper]

Like you said Flimflam, you problem may be a combination of issues which is showing up.

This thread was to answer the logged crash in the "crashlog.txt" of the error:

[VideoPlayer] Failure accessing d3d device

If your crashlog shows no error or it shows a different error this may be a totally

different fault to that one I listed. From what you have posted, it sounds like there

is an issue with a map or maps in the last third of your cycle. Our cycles are normally

around 6 - 8 maps.

do you have a crash log "ENTRY "for the time the crash happened?

viii

[FI_FlimFlam]

yes - every single time. The only difference was the map name indicated if I removed what I mistakenly thought was a bad map.

EDIT: Scratch that - we had 2 full map rotations and then it began again sometime this morning. Again on the 5th map. I have removed the 4th map to see if it is the culprit according to your testing. I'll let you know how it goes.

Edited March 11, 2007 by FI_FlimFlam

[viiiper]

[VideoPlayer] Failure accessing d3d device, ERROR LOG

I have noticed that the Boarders command line in the ZONES.XML

makes a problem of this type when values are all negative

I don't know the exact details but if you push the XY values into a default

-X +Y and +X -Y things work

to prove the point make a std zones.xml for your problem map and insert.

See if it cures this.... mine did and now the problem of not going too negative

so that the mini map overlay is correct.

[tristan73]

I am reeiving the same error messages now on server death2nobodee. Never had these errors before, seems to be random maps in cycle and I have tried changing cycle. Nothing has changed on server other than implementation of new cutsom maps, did anyone get to the bottom of this error?

Deathbringer200