This is not a memory leak really.
Also the ram that you see is not really "dead ram"
I've had about 8-9 processes of javaw.exe and opened a new one.
Normally I should have gotten the blue screen of death if all that ram has been used.
But I didn't.. Which is odd...
Also... People who say that they crash when they click buttons, It happens to me too.
Sometimes I click on the Multiplayer button and I can't connect to any servers. When i go back to mainscreen and click multiplayer, I crash.And btw I'm using a laptop, but I configured my BIOS to use only my ATI graphics card.
Okay, I've tried my hardest to diagnose the issue. And I don't see what could be causing it, it's fine for most users, so it looks like you are the victim of some sort of uncommon Java bug, try contacting them.
Okay, I've tried my hardest to diagnose the issue. And I don't see what could be causing it, it's fine for most users, so it looks like you are the victim of some sort of uncommon Java bug, try contacting them.
Sorry I couldn't help.
I created an account just to reply to this. I am having the exact same problem. These problems, for me are repeatable. javaw.exe is not closing. It will not connect to any multiplayer server. After I click cancel and go to the main menu and click multiplayer again, minecraft freezes. the javaw.exe will not close even if I kill the process. I cannot change my skin, or it is very inconsistent about changing the skin (I don't know if this is related). I can still play single player, but not multiplayer and the process still uses memory. I hope this gets resolved!
Like CraftNinja I've tried and failed to reproduce this.
I can start MC,
Select Multiplayer, Hit Refresh multiple times, Cancel, Re-enter, repeat. Cancel
Select Singleplayer, play, Quit to menu.
Select Multiplayer, Refresh multiple times, Cancel.
Quit.
and minecraft terminates cleanly...
[BTW the skin issue is a completely separate skin server problem]
It only happens when you play for a little bit. Try playing for 5 mins, then try that.
Multiplayer does work, however, I've noticed that if you try to load the multiplayer page, and it fails to check the signal strength of the server, if you click refresh, try to connect or reload the multiplayer page, the game will crash, you can still play SSP after this, but you will probably want to exit cleanly after it fails to check for signal strength.
You might want to make sure that any antivirus you have skips over all the files under your .minecraft/saves directory.
I had a problem the other day after a Quicken upgrade, where it was freaking out while trying to close a data file, back it up, and re-open
it. This was apparently due to my AV jumping in as soon as Quicken wrote to the file during the close. The previous version of quicken didn't have this issue which was a little weird.
Why is this relevant? Well when you shut down MC it has to trigger the server to save all your data, and MC also uses many files. It's possible (with software all things are possible!) that the server save gets hung up while updating multiple files. And then MC may not terminate because it's waiting for the save to complete..
And then this might tie in to those who report single player rollbacks as all being one issue.
Like CraftNinja I've tried and failed to reproduce this.
I can start MC,
Select Multiplayer, Hit Refresh multiple times, Cancel, Re-enter, repeat. Cancel
Select Singleplayer, play, Quit to menu.
Select Multiplayer, Refresh multiple times, Cancel.
Quit.
and minecraft terminates cleanly...
[BTW the skin issue is a completely separate skin server problem]
Regarding the MP refresh freeze, this is what I see in practice:
I play MCSG sometimes, and have all the servers listed. I can join any server as an observer (in effect in a pseudo-creative mode), disconnect, and the server list refreshes fine. If I join the server as a player, play and then disconnect, the server list refresh hangs. Every time. MC is dead at that point. I can attempt to kill the javaw process, which kills the MC session but leaves the process untouched.
I do not know what precisely I am observing, whether it is length of time played on a MP server, whether the creative/survival mode is an issue, or whether it is something else. Whatever the cause, for me this bug it is completely repeatable.
Once again I ask, as a bit of a noob to MC bug reporting, how can we (or can we at all) add this to the list?
I use ClamAV for my antivirus, I occasionally scan my computer, and usually when I do, it's the Downloads directory. I usually go to the same websites, and I try to download free software (not freeware, but free software). I think I'm alright there, but I will do a scan.
You might want to make sure that any antivirus you have skips over all the files under your .minecraft/saves directory.
I had a problem the other day after a Quicken upgrade, where it was freaking out while trying to close a data file, back it up, and re-open
it. This was apparently due to my AV jumping in as soon as Quicken wrote to the file during the close. The previous version of quicken didn't have this issue which was a little weird.
Why is this relevant? Well when you shut down MC it has to trigger the server to save all your data, and MC also uses many files. It's possible (with software all things are possible!) that the server save gets hung up while updating multiple files. And then MC may not terminate because it's waiting for the save to complete..
And then this might tie in to those who report single player rollbacks as all being one issue.
I can correctly exit from a MP server (at least it appears so), would that make this theory less likely? This javaw issue occurs either on the next attempt to refresh a MP list (MC hang) or when quiting MC.
Regarding the MP refresh freeze, this is what I see in practice:
I play MCSG sometimes, and have all the servers listed. I can join any server as an observer (in effect in a pseudo-creative mode), disconnect, and the server list refreshes fine. If I join the server as a player, play and then disconnect, the server list refresh hangs. Every time. MC is dead at that point. I can attempt to kill the javaw process, which kills the MC session but leaves the process untouched.
I do not know what precisely I am observing, whether it is length of time played on a MP server, whether the creative/survival mode is an issue, or whether it is something else. Whatever the cause, for me this bug it is completely repeatable.
Once again I ask, as a bit of a noob to MC bug reporting, how can we (or can we at all) add this to the list?
Couple of other things observed from tonight. The problem is attributable to the length of time on a MP server, in whatever mode. After maybe 5 mins, that's MC's limit. After that, the list will not refresh and the client hangs.
Also, I had thought that logging off cleared the javaw processes, at least it appeared to do so in Task Manager. However, not so, it just hides them. Show Processes from All Users reveals them again, under my local user name. The only way to clear these processes is to reboot.
Prior to reboot, Windows was running at close to max RAM - 90% or so. And will stay there as far as I can see, as Windows seems to be shaving RAM off these processes as it requires, probably paging to swap file (which would explain slowdown in performance I have also observed as the computer starts to get maxed out).
You say the ones where you see an issue are more "graphically enhanced"... by that you mean, you run them at higher settings?
It could be a memory issue, as I've already said, I have absolutely no idea about this issue... but, grasping at straws, running at higher settings will use more memory, and a process will not properly close until it has terminated all memory in use.
Try playing with lower settings, if that makes a difference, then it's time to experiment and see what the culprit setting might be.
Tried the lower 'laptop' settings last night, made no difference. It does seem to me likely to be a hardware specific interaction. Software wise, both x64Win7 and x86 XP desktops show the same bug. I'll do some more digging between the correctly working laptops and the javaw hang desktops.
This really has me stumped as well. I'm currently trying a full un-install of Java. Then a complete re-install. I'll post here if that works but somehow I doubt it :l
Rollback Post to RevisionRollBack
If i helped you, Make sure to click the little "+" on the bottom right of my post.
Java is not the problem guys. I'm pretty sure it has to do with the rushed update of 1.3(leading to bad programming).... and what surprises me is that its not on the bug wiki.
The reason why I don't think it's Java's problem is because it didn't make a difference whether I reinstalled, Used java 6 or 7 or used 32 bit or 64.
I currently have similar issues with multiple instances of javaw.exe running if I don't manually terminate them. This is with a modded mcnostalgia 1.2.5 instance loaded through multimc. Perhaps you could try using something like multimc to test and see if it can kill the minecraft process for you? Just a suggestion.
Java is not the problem guys. I'm pretty sure it has to do with the rushed update of 1.3(leading to bad programming).... and what surprises me is that its not on the bug wiki.
The reason why I don't think it's Java's problem is because it didn't make a difference whether I reinstalled, Used java 6 or 7 or used 32 bit or 64.
Agreed - Java version makes no difference, nor 32/64 bit. I suppose I could try to put it on the bug wiki, but without a crash dump or something specific to analyse, I'm not sure if it will be accepted?
Agreed - Java version makes no difference, nor 32/64 bit. I suppose I could try to put it on the bug wiki, but without a crash dump or something specific to analyse, I'm not sure if it will be accepted?
As of my knowledge it was submitted to the bug wiki before, and it was removed. I'm not 100% sure though because when I look this problem up on Google, the problem was on the description under bug wiki, but when you look at the actual page, it doesn't exist.
It's worth trying through. This seems to be a problem affecting a lot of people.
Also seeing this, but only in MP mode. seems to work fine if i do Single. I tried using process exp (great tool btw!) but the kill process tree option is not available.
I have also re-installed java, (ver 6.35 and 7.7) and both behave this way, while MC 1.2.5 did not do this on either version, but 1.3.2 does. I am using vanilla launcher, but haven't tried running the program in compatibility mode as suggested elsewhere here or in this thread.
I have noticed that as I play more SP mode it tends to take allocated memory from one of my javaw.exe process to make room for the new one. But this solution is almost as bad as restarting in terms of fixing the problem.
Also the ram that you see is not really "dead ram"
I've had about 8-9 processes of javaw.exe and opened a new one.
Normally I should have gotten the blue screen of death if all that ram has been used.
But I didn't.. Which is odd...
Also... People who say that they crash when they click buttons, It happens to me too.
Sometimes I click on the Multiplayer button and I can't connect to any servers. When i go back to mainscreen and click multiplayer, I crash.And btw I'm using a laptop, but I configured my BIOS to use only my ATI graphics card.
I created an account just to reply to this. I am having the exact same problem. These problems, for me are repeatable. javaw.exe is not closing. It will not connect to any multiplayer server. After I click cancel and go to the main menu and click multiplayer again, minecraft freezes. the javaw.exe will not close even if I kill the process. I cannot change my skin, or it is very inconsistent about changing the skin (I don't know if this is related). I can still play single player, but not multiplayer and the process still uses memory. I hope this gets resolved!
I can start MC,
Select Multiplayer, Hit Refresh multiple times, Cancel, Re-enter, repeat. Cancel
Select Singleplayer, play, Quit to menu.
Select Multiplayer, Refresh multiple times, Cancel.
Quit.
and minecraft terminates cleanly...
[BTW the skin issue is a completely separate skin server problem]
It only happens when you play for a little bit. Try playing for 5 mins, then try that.
Multiplayer does work, however, I've noticed that if you try to load the multiplayer page, and it fails to check the signal strength of the server, if you click refresh, try to connect or reload the multiplayer page, the game will crash, you can still play SSP after this, but you will probably want to exit cleanly after it fails to check for signal strength.
Thanks,
Dragos
I had a problem the other day after a Quicken upgrade, where it was freaking out while trying to close a data file, back it up, and re-open
it. This was apparently due to my AV jumping in as soon as Quicken wrote to the file during the close. The previous version of quicken didn't have this issue which was a little weird.
Why is this relevant? Well when you shut down MC it has to trigger the server to save all your data, and MC also uses many files. It's possible (with software all things are possible!) that the server save gets hung up while updating multiple files. And then MC may not terminate because it's waiting for the save to complete..
And then this might tie in to those who report single player rollbacks as all being one issue.
Regarding the MP refresh freeze, this is what I see in practice:
I play MCSG sometimes, and have all the servers listed. I can join any server as an observer (in effect in a pseudo-creative mode), disconnect, and the server list refreshes fine. If I join the server as a player, play and then disconnect, the server list refresh hangs. Every time. MC is dead at that point. I can attempt to kill the javaw process, which kills the MC session but leaves the process untouched.
I do not know what precisely I am observing, whether it is length of time played on a MP server, whether the creative/survival mode is an issue, or whether it is something else. Whatever the cause, for me this bug it is completely repeatable.
Once again I ask, as a bit of a noob to MC bug reporting, how can we (or can we at all) add this to the list?
Thanks,
Dragos
It's Linux software that's used to scan Windows computers. It's also for Windows as well and has a pretty decent database of viruses.
I can correctly exit from a MP server (at least it appears so), would that make this theory less likely? This javaw issue occurs either on the next attempt to refresh a MP list (MC hang) or when quiting MC.
For full disclosure, I use Avast AV.
Couple of other things observed from tonight. The problem is attributable to the length of time on a MP server, in whatever mode. After maybe 5 mins, that's MC's limit. After that, the list will not refresh and the client hangs.
Also, I had thought that logging off cleared the javaw processes, at least it appeared to do so in Task Manager. However, not so, it just hides them. Show Processes from All Users reveals them again, under my local user name. The only way to clear these processes is to reboot.
Prior to reboot, Windows was running at close to max RAM - 90% or so. And will stay there as far as I can see, as Windows seems to be shaving RAM off these processes as it requires, probably paging to swap file (which would explain slowdown in performance I have also observed as the computer starts to get maxed out).
Other java processes that I use do close as normally
It is not lwjgl that's the problem. I updated to latest one and it does nothing. it ususally handles language... so I don't think its even related.
Antivirus is not a problem. I have windows essentials and nothing pops up when i run minecraft.
Repeatable?
Unless you know it happens to you, you can't reproduce it. It has to do with programming code with maybe the the hardware.
HP Envy 14 Beats Edition
Intel Core i7-2670 QM @ 2.2. GHz
6GB RAM
Radeon HD 6630M
Dell Studio XPS 9100
Intel Core i7 930 @2.8Ghz, 4 cores, 8 logical processors
9GB RAM
Radeon HD 5670
Win7 x64 Home Premium
Tried the lower 'laptop' settings last night, made no difference. It does seem to me likely to be a hardware specific interaction. Software wise, both x64Win7 and x86 XP desktops show the same bug. I'll do some more digging between the correctly working laptops and the javaw hang desktops.
The reason why I don't think it's Java's problem is because it didn't make a difference whether I reinstalled, Used java 6 or 7 or used 32 bit or 64.
Agreed - Java version makes no difference, nor 32/64 bit. I suppose I could try to put it on the bug wiki, but without a crash dump or something specific to analyse, I'm not sure if it will be accepted?
As of my knowledge it was submitted to the bug wiki before, and it was removed. I'm not 100% sure though because when I look this problem up on Google, the problem was on the description under bug wiki, but when you look at the actual page, it doesn't exist.
It's worth trying through. This seems to be a problem affecting a lot of people.
I have also re-installed java, (ver 6.35 and 7.7) and both behave this way, while MC 1.2.5 did not do this on either version, but 1.3.2 does. I am using vanilla launcher, but haven't tried running the program in compatibility mode as suggested elsewhere here or in
this thread.
I have noticed that as I play more SP mode it tends to take allocated memory from one of my javaw.exe process to make room for the new one. But this solution is almost as bad as restarting in terms of fixing the problem.
In case you didn't read my other post about what DOESNT work
And yes I've tried programs like Process Explorer and many other programs that were designed to "kill processes."