SphereServer BugTracker - SphereServer |
View Issue Details |
|
ID | Project | Category | View Status | Date Submitted | Last Update |
0001891 | SphereServer | executable - windows build | public | 17-01-11 12:31 | 24-01-11 15:27 |
|
Reporter | Terrikate | |
Assigned To | MrSugarCube | |
Priority | normal | Severity | block | Reproducibility | have not tried |
Status | closed | Resolution | duplicate | |
Platform | | OS | | OS Version | |
Product Version | | |
Target Version | | Fixed in Version | | |
Nightly Version | Automated (specify build number) |
Experimental Flags | DiagonalWalkCheck, New_Triggers, Script_Profiler, Size_Optimise, UseNetworkMulti |
Option Flags | Command_Sysmsgs, FileCommands, Buffs |
Internal Build Number | 1395 |
|
Summary | 0001891: Compress failed with error -4 when generating gump. Using old packet. |
Description | Yesterday I switched to the latest version of the emulator (I was using 3 January, I was wrong) and I found the error in the description in less than 24 hours on two occasions.
In the first case with UseExtraBuffer enabled server still worked but people could not connect, no saves were made.
The second error after 3 minutes the server crashed.
Sphere reports (crash dumps) are all 0KB.
All players are using UO: 2D Client SA 07082 |
Steps To Reproduce | |
Additional Information | |
Tags | No tags attached. |
Relationships | related to | 0001868 | resolved | MrSugarCube | Sphere has a major memory leak |
|
Attached Files | logs.zip (258,824) 17-01-11 12:31 http://tracker.sphere.torfo.org/bugs/file_download.php?file_id=51&type=bug sphere_block_again.jpg (431,716) 17-01-11 22:05 http://tracker.sphere.torfo.org/bugs/file_download.php?file_id=52&type=bug |
|
Issue History |
Date Modified | Username | Field | Change |
17-01-11 12:31 | Terrikate | New Issue | |
17-01-11 12:31 | Terrikate | File Added: logs.zip | |
17-01-11 13:31 | MrSugarCube | Note Added: 0000840 | |
17-01-11 13:42 | Terrikate | Note Added: 0000841 | |
17-01-11 22:05 | Terrikate | File Added: sphere_block_again.jpg | |
17-01-11 22:09 | Terrikate | Note Added: 0000842 | |
17-01-11 23:33 | MrSugarCube | Note Added: 0000843 | |
17-01-11 23:59 | Terrikate | Note Added: 0000844 | |
19-01-11 16:19 | Terrikate | Note Added: 0000845 | |
19-01-11 17:00 | MrSugarCube | Relationship added | related to 0001868 |
19-01-11 18:03 | MrSugarCube | Note Added: 0000846 | |
20-01-11 13:17 | Terrikate | Note Added: 0000850 | |
22-01-11 02:22 | MrSugarCube | Note Added: 0000853 | |
24-01-11 00:24 | Terrikate | Note Added: 0000863 | |
24-01-11 00:38 | Terrikate | Note Added: 0000864 | |
24-01-11 00:52 | MrSugarCube | Note Added: 0000865 | |
24-01-11 01:41 | Terrikate | Note Added: 0000867 | |
24-01-11 02:10 | Terrikate | Note Edited: 0000867 | bug_revision_view_page.php?bugnote_id=0000867#r97 |
24-01-11 15:18 | Terrikate | Note Added: 0000868 | |
24-01-11 15:27 | MrSugarCube | Status | new => closed |
24-01-11 15:27 | MrSugarCube | Assigned To | => MrSugarCube |
24-01-11 15:27 | MrSugarCube | Resolution | open => duplicate |
Notes |
|
|
The error code indicates that there was not enough memory available to compress the packet.
Did you notice that memory usage was particularly high at the time this happened? |
|
|
|
No, but now sphere has been to use 1065MB when it usually did not exceed the 950MB.
If you want to check one more thing ... |
|
|
|
I have been blocked again (I'm back to a previous version). If you need some more information this time I made a dump of the entire process before closing. Information about 2GB. About 230MB compressed. |
|
|
|
The previous version you have switched back to is from 3rd January 2011 (1385?), and doesn't have this problem? |
|
|
|
|
|
|
After further testing version 3 January and 23 December the problem persists ... so I gather it's just my thing. As you mention the ruling party is more hardware than software, right?.
But wonder if hardware does not strike the typical blue screen and restart the computer. I'll see automatic updates have been made recently. Surprise me was just a RAM problem.
If you know of anything else I'm all ears.
Thanks. |
|
|
|
It won't be a hardware problem (at least, there's no need to check your RAM or anything :)). Based on the task manager image it appears that the problem is caused by Sphere (for some reason) trying to use more than 2GB of RAM and Windows does not allow this by default.
This could indicate that Sphere has a memory leak or possibly that it is trying to do something that needs this much memory. It could be worth us adding support for using additional memory (3GB or 4GB max depending on 32-bit or 64-bit), although if the problem is a memory leak then it will only buy some extra time before the issue arises.
I don't know what the side effects will be (if any) when enabling support for >2GB memory, but there's a guide at http://www.techpowerup.com/forums/showthread.php?t=112556 [^] if you want to try it yourself at your own risk. |
|
|
|
Thx.
Took some time without controlling the use of RAM and so, if before remained constant below 1 GB are now gradually increases without releasing anything at any time.
I told the process use more than 2GB of RAM so we'll know in a moment if it serves as a temporary fix or not to avoid having to restart the server every day. |
|
|
|
Since issue 0001868 has been fixed (hopefully), it might be worth trying out the next automated build (version 1401) to check if the situation has improved at all.
If you have TooltipCache=0 in your sphere.ini then it is possibly the same issue that you're seeing here. |
|
|
|
Yesterday's version continues with the same problem and the last compiled is unplayable, much lag. The server gets to 100% CPU. |
|
|
|
Wait for you that the issue is more rare than I thought. I'll tell you that maybe the problem is the equipment itself. It was back to a previous version and is just as bad. |
|
|
|
Build 1401 didn't fully address issue 0001868 so you'll be better off checking tonights automated build instead (version 1402), or perhaps enable TooltipCache to eliminate that as a possible cause. |
|
|
(0000867)
|
Terrikate
|
24-01-11 01:41
(edited on: 24-01-11 02:10) |
|
I tried both, #1401 and #1402 (for few minutes). The issue is that It got 100% cpu after starting the server. It rarely happens. Returning to yesterday's version, still going on but after making a massive disconnect and was well ... Tomorrow I'll tell you how everything went.
|
|
|
|
|