aMule Forum
English => aMule Help => Topic started by: Ezeltje on March 20, 2006, 12:33:15 AM
-
I downloaded v2.1.1 and compiled it this afternoon. After running ~3 1/2 hours, amule was using 1148m virtual, 784m shared memory, according to top. If there is any way I can help track this problem down, please let me know.
-
amule --version
-
$ ./amule --version
aMule 2.1.1 using wxGTK2 v2.6.2 (Unicoded) (OS: Linux)
-
Do you use EC?
-
You mean External Connections? Yes, for quite some time.
-
Update: I've had 2.1.1 running now for ~7 hours, and memory usage appears to have stabilized at 106m RES and 147m VIRT. Apparently, the mem leak only happens under very specific circumstances...
-
There is a potentially huge memleak on EC I fixed yesterday, will be ok on 2.1.2 (which would be.... soon)
-
Originally posted by Kry
There is a potentially huge memleak on EC I fixed yesterday, will be ok on 2.1.2 (which would be.... soon)
Has this made it into the CVS version? I am running the 1 April tarball and just now the memory leak once again completely choked my system; it became unresponsive and I had to hit the reset button. Although I had EC enabled, I wasn't using it, but for the time being I have disabled it.
I really like amule -- but not when it nukes my system. ;(
-
aMule can't nuke your system. Which wx are you using?
-
Never mind. it appears that it was a CUPS program called footmatic-rip that was choking my system. Problem was that I couldn't open top in time to find out what the culprit was... Sorry for the false alarm.
-
Told ya, told ya....
-
My system spontaneously started thrashing again this morning and this time it did in fact turn out to be caused by amule. Although the system became unresponsive, I was finally able to shutdown X with ctrl-alt-backspace and log into a failsafe terminal. Here is the output from top which I saved to disk:
$ cat toplog.txt
top - 12:19:25 up 1 day, 11:41, 3 users, load average: 1.69, 9.37, 12.31
Tasks: 95 total, 1 running, 94 sleeping, 0 stopped, 0 zombie
Cpu(s): 20.6% us, 1.9% sy, 0.0% ni, 0.0% id, 77.5% wa, 0.0% hi, 0.0% si
Mem: 1035176k total, 1020624k used, 14552k free, 620k buffers
Swap: 1020116k total, 534784k used, 485332k free, 10812k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3049 jrb 18 0 1339m 899m 2492 D 21.1 89.0 262:51.90 amule
157 root 15 0 0 0 0 S 0.3 0.0 0:37.54 kswapd0
2347 root 15 0 0 0 0 S 0.3 0.0 0:10.13 nfsd
1 root 16 0 1992 104 80 S 0.0 0.0 0:01.22 init
2 root 34 19 0 0 0 S 0.0 0.0 0:00.48 ksoftirqd/0
3 root RT 0 0 0 0 S 0.0 0.0 0:00.00 watchdog/0
4 root 10 -5 0 0 0 S 0.0 0.0 0:00.32 events/0
5 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 khelper
6 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 kthread
8 root 10 -5 0 0 0 S 0.0 0.0 0:00.62 kblockd/0
9 root 20 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid
100 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 khubd
158 root 18 -5 0 0 0 S 0.0 0.0 0:00.00 aio/0
245 root 10 -5 0 0 0 S 0.0 0.0 0:00.02 kseriod
320 root 11 -5 0 0 0 S 0.0 0.0 0:00.00 kpsmoused
336 root 10 -5 0 0 0 S 0.0 0.0 0:00.02 ata/0
340 root 11 -5 0 0 0 S 0.0 0.0 0:00.00 scsi_eh_0
As you can see, amule was still running and using a catastrophic amount of memory, nearly 900M on a 1 GB machine. For some reason it had not responded to the Gnome logout and I could only terminate it with killall -9.
FWIW: amule had been running about 36 hours. Yes, I am sharing a lot of files.
At this point, I don't think it has anything to do with ext connections; I had that function disabled and amuleweb was not running. This is the 1 April CVS with wxGTK-2.6.2-5.fc5.
Let me know if I can supply any additional information...
-
This time amule ran fine for about three days, then suddenly started gobbling up memory. I captured the top output in a terminal window on my laptop:
$ top
top - 20:05:51 up 4 days, 19:28, 5 users, load average: 14.24, 7.73, 4.17
Tasks: 168 total, 1 running, 166 sleeping, 0 stopped, 1 zombie
Cpu(s): 1.6% us, 10.9% sy, 0.0% ni, 0.0% id, 87.4% wa, 0.0% hi, 0.0% si
Mem: 1035176k total, 1022996k used, 12180k free, 548k buffers
Swap: 1020116k total, 1020116k used, 0k free, 21536k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
22470 jrb 17 0 1218m 682m 2436 D 1.0 67.5 611:06.59 amule
32682 jrb 16 0 215m 63m 3412 D 0.5 6.3 0:28.74 firefox-bin
23291 jrb 16 0 299m 49m 1992 D 0.5 4.9 10:23.91 thunderbird-bin
22155 root 16 0 160m 30m 1788 D 1.3 3.0 905:14.19 Xorg
[...]
Although the desktop box became unresponsive, I fortunately got to my laptop in time to kill amule via this terminal window.
I am the only one having this problem???
-
Seems so. weird.
-
Could there possibly be something in the 2.1 kad code? As I mentioned earlier in this thread, I never had this problem with 2.03, which I ran for many weeks, sometimes months at a time.
-
08:48:15 up 3 days, 6:21, 1 user, load average: 0.95, 0.93, 0.91
67 processes: 64 sleeping, 3 running, 0 zombie, 0 stopped
CPU0 states: 5.1% user 5.0% system 0.0% nice 0.0% iowait 89.2% idle
CPU1 states: 2.2% user 28.4% system 0.0% nice 0.0% iowait 68.2% idle
Mem: 2032076k av, 2019084k used, 12992k free, 0k shrd, 62884k buff
1431248k actv, 172832k in_d, 43744k in_c
Swap: 1052248k av, 196212k used, 856036k free 1626792k cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
7279 admin 15 0 53224 51M 3252 R 9.7 2.6 21:46 0 amuled
I'm running 100 files, phew takes a lot of cpu resources, but better than torrent
running amule 2.1.1-2 w/ gtk 2.6.3
daemon better than gui then...
-
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 9.1 12.1 196540 125496 pts/6 Sl 21:54 3:54 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 10.8 12.1 196540 125960 pts/6 Sl Apr17 16:06 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 10.8 12.1 196540 125968 pts/6 Sl Apr17 16:25 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 11.0 12.1 196540 125968 pts/6 Sl Apr17 18:08 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 17.2 12.1 196540 125992 pts/6 Sl Apr17 97:22 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 17.5 12.1 196540 126068 pts/6 Dl Apr17 101:06 trunk/src/amule --only-chuck-norris-would-stop-me
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
myuser 18921 17.5 12.1 196540 126068 pts/6 Rl Apr17 101:13 trunk/src/amule --only-chuck-norris-would-stop-me
...
aMule CVS rev. 6702 has been running for 9:45 h
amuleuser is connnnected to xxxxxxxxxxxxxxx [XXX.XXX.XXX.XXX:XXXX] with HighID | Kad: ok
Total Download: 378.82 Gb, Upload: 193.71 Gb
Session Download: 1.42 Gb, Upload: 628.63 Mb
Download: 209.2 kB/s, Upload: 24.0 kB/s
Sharing: 1190 file(s), Clients on queue: 4999
This is the latest CVS code, aMule monolithic. Memory usage seems stable. 50 files in dl queue.
-
Ok, back again. I am now running today's CVS. How do you get those amule stats on the command line?
-
Use a program named cas. You must configure aMule with --enable-cas or something like this.
-
Here is my cas output:
aMule CVS Thu May 4 07:02:02 CEST 2006 has been running for 1:14 h
amuleuser is connnnected to XXXXXXXXXX [xxxxxxxxxxxxx] with HighID | Kad: ok
Total Download: 652.67 Gb, Upload: 2.05 Tb
Session Download: 37.35 Mb, Upload: 334.23 Mb
Download: 6.4 kB/s, Upload: 80.0 kB/s
Sharing: 1371 file(s), Clients on queue: 4817
Any other information I can supply you with?
-
Well, cas does not give us any interesting information, I had posted it only to show my uptime.
You said you have problems with memory leaks. I assume you use CVS code, right? Post here:
$ amule --version
And with aMule running, with one hour intervals, run
$ ps u -C amule
So that we can compare the memory usage report.
Cheers!
-
seems same problem.
Starts with 10.5 % mem usage and when it reaches +- 60% i can't preview files
If i close and restart amule then all is ok.
suse system. Amule and wxGTK2 compiled as says in amule wiki "compile and install amule from source"
amule --version :
aMule 2.1.1 using wxGTK2 v2.6.3 (Unicoded) (OS: Linux)
"ps u -C amule" with a 30 minutes interval :
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.2 37.6 418948 389656 ? Sl May04 265:09 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.2 37.8 421648 392436 ? Sl May04 269:12 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.2 38.7 430844 401348 ? Sl May04 278:56 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.2 38.8 431836 402516 ? Rl May04 283:05 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.2 39.1 434892 405524 ? Sl May04 287:56 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.1 39.3 437536 408104 ? Sl May04 292:29 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.1 39.7 441256 412112 ? Sl May04 296:54 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.1 40.1 444980 415836 ? Sl May04 301:05 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.0 40.5 448820 419820 ? Sl May04 304:58 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.0 40.9 453528 424520 ? Sl May04 309:20 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.0 41.3 458016 428524 ? Sl May04 313:31 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 16.0 41.8 462756 433468 ? Sl May04 318:07 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.9 42.1 466188 436884 ? Sl May04 322:18 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.9 42.5 470356 440924 ? Sl May04 326:41 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.9 42.8 473868 444504 ? Sl May04 330:38 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.8 43.2 478060 448612 ? Sl May04 334:29 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.8 43.7 482364 453308 ? Sl May04 338:39 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.8 44.2 487152 458008 ? Sl May04 343:16 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.8 44.5 490688 461680 ? Sl May04 347:47 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 44.8 494292 465192 ? Sl May04 351:48 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 45.3 499528 470156 ? Sl May04 356:17 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 45.7 503100 473812 ? Sl May04 360:36 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 46.1 507948 478660 ? Sl May04 365:03 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 46.6 513140 483816 ? Sl May04 369:44 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 47.1 518012 488888 ? Sl May04 374:19 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 47.6 522692 493396 ? Sl May04 379:13 amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
basura 30923 15.7 48.0 526816 497728 ? Sl May04 383:37 amule
any help?
-
That's really weird.
Wait, do you use EC?
-
EC? --> external conections?
No, i do not use external conections
-
Yeah, definetly sucking memory...
Can you compile and try a CVS tarball?
-
Hi all,
Since compiling the 4 May CVS, I have not seen the memory leak. In the beginning, it crashed a couple of times after running ~24 hours, but I now have an uptime of six days:
$ cas
aMule CVS Thu May 4 07:02:02 CEST 2006 has been running for 6 D 00 h
amule-user is connnnected to xxxxxxxxxxx [XXXXXXXXXXXXX] with HighID | Kad: ok
Total Download: 658.58 Gb, Upload: 2.12 Tb
Session Download: 1.41 Gb, Upload: 38.67 Gb
Download: 0.0 kB/s, Upload: 79.9 kB/s
Sharing: 1383 file(s), Clients on queue: 4997
It appears the problem has been fixed.
Thanks everyone. aMule is awesome!
-
Alas, I encountered another memory leak today, running the 2.1.2 code, with EC enabled.
Tue Jun 6 17:25:02 CEST 2006
aMule 2.1.2 has been running for 1 D 00 h
amule-user is connnnected to XXXXXXXXXXXX [xxxxxxxxxxxxxx] with HighID | Kad: ok
Total Download: 668.75 Gb, Upload: 2.22 Tb
Session Download: 2.10 Gb, Upload: 6.76 Gb
Download: 14.6 kB/s, Upload: 69.8 kB/s
Sharing: 1417 file(s), Clients on queue: 5000
At Tue Jun 6 17:25:02 CEST 2006:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xxxxx 32372 19.1 11.0 238288 114760 ? Sl Jun05 286:00 /usr/local/bin/amule
Five minutes later:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xxxxx 32372 19.3 70.9 1094492 734064 ? Rl Jun05 289:37 /usr/local/bin/amule
top indicated 773M resident memory, 1240M shared before I killed it.
Please don't shoot the messenger :(
-
Ezeltje,
you said something about 2.0.3. How you comment this ?
This is aMule 2.0.3 using wxGTK2 v2.6.1 (Unicoded) based on eMule.
Running on Linux 2.6.12-10-386 i686
top - 20:30:12 up 11 days, 2 min, 2 users, load average: 0.09, 0.19, 0.17
Tasks: 89 total, 1 running, 88 sleeping, 0 stopped, 0 zombie
Cpu(s): 4.3% us, 1.0% sy, 0.0% ni, 94.4% id, 0.0% wa, 0.0% hi, 0.3% si
Mem: 776552k total, 767108k used, 9444k free, 11460k buffers
Swap: 499928k total, 418684k used, 81244k free, 149468k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
28310 xxxxxx 16 0 838m 441m 11m S 3.3 58.2 334:50.32 amule
27984 root 15 0 123m 35m 7860 S 1.3 4.7 195:57.20 Xorg
28146 xxxxxx 16 0 104m 37m 13m S 0.3 4.9 5:25.08 nautilus
Keep on 8o
-
walkin: There's a huge memleak in wx < 2.6.2 and I'm not sure if the workaround in amule was added before or after 2.0.3. Try upgrading wx & amule.
-
The mem leak snuck up on me again a few days ago (I am running the 20060724 CVS), so I wrote a little script to prevent amule from choking my system when the memleak occurs:
#!/bin/bash
#
amule_max_mem="500000"
sleep_interval="60"
while :
do
amule_virtual_mem=`ps u -C amule | tail -1 | awk '{print $5}'`
if [ $amule_virtual_mem -gt $amule_max_mem ]; then
killall amule
fi
sleep $sleep_interval
done
This script checks once a minute how much virtual memory amule is using and kills if it starts using more than 500MB (on my system, it tends to use around 280MB under normal circumstances).
If anyone else is encountering the mem leak, I hope they will find this script useful.
-
Originally posted by thedude0001
walkin: There's a huge memleak in wx < 2.6.2 and I'm not sure if the workaround in amule was added before or after 2.0.3. Try upgrading wx & amule.
I think amule v2.1.3 has the memory leak too.
The wx I use is 2.7.0-1.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1193 zhengda 25 0 610m 400m 9188 R 33.0 80.9 10:42.50 amule
4194 root 6 -10 84436 18m 4096 S 27.0 3.7 30:46.38 XFree86
4596 zhengda 16 0 72396 9224 6292 R 17.0 1.8 2:58.10 gnome-terminal
-
Originally posted by icoming
I think amule v2.1.3 has the memory leak too.
Yep, I still encounter it every few days...
-
How to disable external conections?
Actually, I don't know whether it has been enabled, but I want a try.
-
Preferences > Remote Controls > Accept external connections
or
edit amule.conf, in the section [ExternalConnect] set AcceptExternalConnections=0
-
Then I haven't enabled it.
aMule must have some other functions which cause the memory leak.
Which latest version of aMule has not been report having the memory leak?
-
Originally posted by icoming
aMule must have some other functions which cause the memory leak.
Which latest version of aMule has not been report having the memory leak?
v.2.0. It appears to have been introduced in v.2.1.1. I thought the addition of kad might be responsible for it, but I encounter the memleak even with kad turned off.
-
I think aMule 2.1.3 has another problem.
aMule runs for a while, and it uses the most resources in my computer.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
12368 zhengda 25 0 330m 247m 16m R 99.2 50.0 18:15.34 amule
4587 root 5 -10 56368 34m 7428 S 0.3 7.0 25:22.79 XFree86
-
I've noticed my system becoming really sluggish and 100% out of memory as shown by GNOME
System Monitor. Never knew what program was doing it til today. I was able to ssh in from another
machine and run "top", showing 1400+ Mbyte amule process. When I created a new swapfile and
did "swapon" to it, within seconds amule had eaten another 64 megs. I added two more 128 meg
swapfiles and those got eaten too. But I was able to kill -ILL the amule process, hoping to get
a core file or something that would allow this to be debugged.
Unfortunately, amule contains some silly signal catcher which prevents it from dumping core --
instead I got this output:
$ kill -ILL %3
$
--------------------------------------------------------------------------------
A fatal error has occurred and aMule has crashed.
Please assist us in fixing this problem by posting the backtrace below in our
'aMule Crashes' forum and include as much information as possible regarding the
circumstances of this crash. The forum is located here:
http://forum.amule.org/board.php?boardid=67
If possible, please try to generate a real backtrace of this crash:
http://www.amule.org/wiki/index.php/Backtraces
----------------------------=| BACKTRACE FOLLOWS: |=----------------------------
Current version is: aMule 2.1.3 using wxGTK2 v2.6.3 (Unicoded)
Running on: Linux 2.6.11-1-k7-smp i686
[then a lot of BEEPs, as if control-G was being output many times...]
$ jobs
[3]+ Running amule & (wd: ~)
$ kill -ILL %3
$
--------------------------------------------------------------------------------
A fatal error has occurred and aMule has crashed.
Please assist us in fixing this problem by posting the backtrace below in our
'aMule Crashes' forum and include as much information as possible regarding the
circumstances of this crash. The forum is located here:
http://forum.amule.org/board.php?boardid=67
If possible, please try to generate a real backtrace of this crash:
http://www.amule.org/wiki/index.php/Backtraces
----------------------------=| BACKTRACE FOLLOWS: |=----------------------------
Current version is: aMule 2.1.3 using wxGTK2 v2.6.3 (Unicoded)
Running on: Linux 2.6.11-1-k7-smp i686
[2] ?? in amule [0x8083d46]
[3] wxFatalSignalHandler in /usr/lib/libwx_baseu-2.6.so.0[0xb75b66c8]
[4] ?? in [0xffffe420]
[5] wxMilliSleep(unsigned long) in /usr/lib/libwx_baseu-2.6.so.0[0xb75b4705]
[6] wxThread::Sleep(unsigned long) in /usr/lib/libwx_baseu-2.6.so.0[0xb75ac84f]
[7] ?? in amule [0x8117df4]
[8] wxThreadInternal::PthreadStart(wxThread*) in /usr/lib/libwx_baseu-2.6.so.0[0xb75ac1c6]
[9] wxPthreadStart in /usr/lib/libwx_baseu-2.6.so.0[0xb75ac03f]
[10] ?? in /lib/tls/libpthread.so.0 [0xb7fcab63]
[11] __clone in /lib/tls/libc.so.6[0xb734d18a]
--------------------------------------------------------------------------------
[3]+ Aborted amule (wd: ~)
Note that I had to kill it twice to get any meaningful output.
This bug only happens to me every few weeks, it seems. I use both kad and ed2k.
I had external connections enabled, but wasn't using it (it was there for running amule-daemon,
but I discovered amule-daemon didn't support double clicking on a file to see what peers have
it, which I use a lot, so I stopped using it).
-
Originally posted by regarding
This bug only happens to me every few weeks
You are lucky then. The longest uptime I have seen since 2.1 is 12 days. Sometimes the memleak pops up after just 4 or 5 hours but on average it is about two days.
-
Hi all,
I hate to nag, but is there any chance this bug will ever be fixed?
I just encountered it again this evening; in the course of about eight minutes, aMule went from around 100M/240M to 680M/1648M (real/virtual) memory use, thereby bringing my system to a halt and requiring a reboot. These are entries from a log updated every two minutes:
Fri Dec 8 00:42:02 CET 2006
aMule CVS Sun Dec 3 07:01:55 CET 2006 has been running for 4:21 h
xxxxxxx is connnnected to XXXXXXXXXXXXXX [x.x.x.x:4242] with HighID | Kad: ok
Total Download: 1.32 Tb, Upload: 3.16 Tb
Session Download: 860.17 Mb, Upload: 424.04 Mb
Download: 53.6 kB/s, Upload: 27.1 kB/s
Sharing: 652 file(s), Clients on queue: 1705
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xyz 7868 7.7 9.9 242352 103356 ? RNl Dec07 20:21 amule --even-if-lfroen-complains-this-will-stay
[...]
Fri Dec 8 00:50:05 CET 2006
aMule CVS Sun Dec 3 07:01:55 CET 2006 has been running for 4:22 h
xxxxxxx is connnnected to XXXXXXXXXXXXXX [x.x.x.x:4242] with HighID | Kad: ok
Total Download: 1.32 Tb, Upload: 3.16 Tb
Session Download: 862.55 Mb, Upload: 425.29 Mb
Download: 58.3 kB/s, Upload: 31.7 kB/s
Sharing: 652 file(s), Clients on queue: 1701
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xyz 7868 9.3 66.4 1648336 686984 ? DNl Dec07 25:04 amule --even-if-lfroen-complains-this-will-stay
As I mentioned before in this thread, I first encountered this memory leak in January with v2.1.1. I really hope this gets fixed eventually.
$ rpm -qa | grep wx
wxGTK-gl-2.6.3-2.6.3.2.3.fc6
wxGTK-devel-2.6.3-2.6.3.2.3.fc6
wxGTK-2.6.3-2.6.3.2.3.fc6
-
...and again tonight: after an uptime of 1 hour, 43 minutes X( X( X(
aMule CVS Sun Dec 3 07:01:55 CET 2006 has been running for 1:43 h
xxxxxxx is connnnected to XXXXXXXXXXXXXX [x.x.x.x:4242] with HighID | Kad: ok
Total Download: 1.33 Tb, Upload: 3.16 Tb
Session Download: 326.13 Mb, Upload: 172.89 Mb
Download: 50.7 kB/s, Upload: 39.7 kB/s
Sharing: 657 file(s), Clients on queue: 2332
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xyz 3458 13.1 82.1 1713464 849312 ? DNl 00:01 14:45 amule --even-if-lfroen-complains-this-will-stay
-
Ezeltje,
aMule runs in my computer for days without noticeable increase in memory usage, and believe me, I do stress aMule. I stress it to the point that I have reached a limit of open files in my system, when aMule starts to download like 800 files, it starts to fail. My next task will be to test a patch from a user that claims to only open the files when they are needed, thus increasing that limit a lot. My memory consumption is always around 10-13%. I have 512 mega of ram. I share 1580 files.
aMule CVS rev. 6988 has been running for 4:18 h
xxxxxxxxxx is connnnected to xxxxxxxxxxxxxxxx [xxx.xxx.xxx.xxx:xxxx] with HighID | Kad: ok
Total Download: 948.26 Gb, Upload: 345.63 Gb
Session Download: 491.68 Mb, Upload: 328.31 Mb
Download: 58.0 kB/s, Upload: 23.7 kB/s
Sharing: 1580 file(s), Clients on queue: 5000
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xxxxxx 31592 21.4 10.0 263472 103576 pts/3 Dl+ Dec08 55:30 /home/xxxxxx/amule --even-if-lfroen-complains-this-will-
Maybe your leak is somewhere else, maybe in wxGTK or gtk itself. I have heard that there could be a problem between two certain versions of wxGTK and gtk.
-
Originally posted by phoenix
aMule runs in my computer for days without noticeable increase in memory usage, and believe me, I do stress aMule. I stress it to the point that I have reached a limit of open files in my system, when aMule starts to download like 800 files, it starts to fail.
Wow, reminds me of DOS days (<1990). Remember the default FILES=20 in config.sys? ;)
I had also thought I was stressing aMule too much, sharing 1600 files (~600Gb), so I experimented for a few days with just sharing my current downloads, but it didn't make much difference in stability.
Maybe your leak is somewhere else, maybe in wxGTK or gtk itself. I have heard that there could be a problem between two certain versions of wxGTK and gtk.
I am going to use amuled for the next week or so. If your hypothesis is correct, then I shouldn't see the leak. wx/gtk is only used for GUI stuff, right?
Later, I will also try running amule under Ubuntu to see if that makes any difference.
Thanks for the reply.
-
Originally posted by Ezeltje
I am going to use amuled for the next week or so. If your hypothesis is correct, then I shouldn't see the leak. wx/gtk is only used for GUI stuff, right?
Later, I will also try running amule under Ubuntu to see if that makes any difference.
Thanks for the reply.
Excelent idea! Please report your findings.
-
I now have an uptime with amuled of 21 days, so I think I can safely say the mem leak is not in the core:
$ cas
aMule CVS Sat Dec 16 07:01:50 CET 2006 has been running for 21 D 14 h
xxxxxxxxxx is connnnected to xxxxxxxxxxxxxxxx [xxx.xxx.xxx.xxx:xxxx] with HighID | Kad: off
Total Download: 1.40 Tb, Upload: 3.31 Tb
Session Download: 11.52 Gb, Upload: 71.57 Gb
Download: 2.8 kB/s, Upload: 45.0 kB/s
Sharing: 239 file(s), Clients on queue: 763
(BTW, I don't know why cas reports "Kad: off"; kad searches with amulecmd work fine.)
Virtual memory use has crept up to 637 MB, which seems like a lot:
$ ps u -C amuled
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xyz 3196 1.4 6.6 637488 68712 ? Ssl 2006 465:38 amuled -f
but resident memory is only 68 MB and at least it hasn't gone haywire.
At this point, I am considering running aMule monolithic under Ubuntu for awhile. Rather than copy ~/.amule to my laptop (../incomplete = ~15 Gb) I am thinking of just mounting it as an NFS share. Is this a Bad Idea???
Another option might be to download and compile v2.8 of wxGTK on my FC6 box to see if it fixes the memleak under Fedora, but only if it won't trash vlc, my main media player. Thoughts anyone?
-
Hi phoenix,
Originally posted by phoenix
Ezeltje,
aMule runs in my computer for days without noticeable increase in memory usage, and believe me, I do stress aMule. I stress it to the point that I have reached a limit of open files in my system, when aMule starts to download like 800 files, it starts to fail. My next task will be to test a patch from a user that claims to only open the files when they are needed, thus increasing that limit a lot. My memory consumption is always around 10-13%. I have 512 mega of ram. I share 1580 files.
aMule CVS rev. 6988 has been running for 4:18 h
xxxxxxxxxx is connnnected to xxxxxxxxxxxxxxxx [xxx.xxx.xxx.xxx:xxxx] with HighID | Kad: ok
Total Download: 948.26 Gb, Upload: 345.63 Gb
Session Download: 491.68 Mb, Upload: 328.31 Mb
Download: 58.0 kB/s, Upload: 23.7 kB/s
Sharing: 1580 file(s), Clients on queue: 5000
$ ps u -C amule
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
xxxxxx 31592 21.4 10.0 263472 103576 pts/3 Dl+ Dec08 55:30 /home/xxxxxx/amule --even-if-lfroen-complains-this-will-
Maybe your leak is somewhere else, maybe in wxGTK or gtk itself. I have heard that there could be a problem between two certain versions of wxGTK and gtk.
Can you tell me which configure options did you use for wxGTK and aMule CVS? Have you compiled GTK too?
Thank you
-
Originally posted by niels
Hi phoenix,
Can you tell me which configure options did you use for wxGTK and aMule CVS? Have you compiled GTK too?
Thank you
Sure, but please notice that I am not the usual user. I always use debug flags, and I have more than one wxWidget version installed in my system. If you have any doubts, please ask.
I have not compiled gtk.
For wxGTK I do:
$ ./configure --enable-debug --disable-optimise --enable-debug_flag --enable-debug_info --enable-debug_gdb --with-opengl --enable-gtk2 --enable-unicode --enable-largefile --prefix=/home/myuser/usr/local/wxWidgets-2.8.0 && make && sudo make install
For aMule I will correspondingly do (notice the --prefix parameter above):
$ ./autogen.sh && ./configure --with-denoise-level=3 --enable-debug --disable-optimize --enable-verbose --enable-webserver --enable-webservergui --enable-amulecmd --enable-amulecmdgui --enable-ccache --enable-amule-daemon --with-wx-config=/home/myuser/usr/local/wxWidgets-2.8.0/bin/wx-config --enable-cas --enable-amule-gui && LD_LIBRARY_PATH=/home/myuser/usr/local/wxWidgets-2.8.0/lib/ make
If you want, you can add a --prefix to the amule compilation line too. Notice also that I do not mess with the system directories, I always install stuff in my own home directory, unless I use a system package like rpm or deb. I don't eve run make install on aMule, I run it from the command line in the directory it has been compiled like this (trunk, in this case):
$ LD_LIBRARY_PATH=/home/myuser/usr/local/wxWidgets-2.8.0/ LANG=en_US.UTF-8 trunk/src/amule --even-if-lfroen-complains-this-will-stay
As you can see, this can be simplified a lot. You probably do not need to build all the stuff like I do. Also, if you install wxGTK in a directory listed in ld.so.conf, you dont need the LD_LIBRARY_PATH stuff. Also, if your LANG is set, you dont need it too.
Good luck!
-
Thank you phoenix.
I'll test the aMule compiled with your instructions for a few days.
I have the same memory leak than Ezeltje, but i installed the aMule rpm for openSuse 10.2 from http://gw78.ja-nee.de/suse/10.2/
I don't know if the configure options will fix the problem.
I'll post the results.
Thanks again.
-
Niels,
I will be very interested in hearing about your experiences.
One thing: have you also compiled wxGTK? If so, can you run vlc to see whether it works under your configuration? It is the only other wx-based program I use, and it would be really inconvenient if a custom wx install nuked it.
TIA
-
Hi Ezeltje,
If you have problem with vlc and two versions of wxGTK, one installed in your home directory and another installed in the system directories, you can do one thing:
Instead use the path of vlc (for example: /usr/bin/vlc) in aMule -> Preferences -> Directories -> Video Player, use the path to a script (for example: /home/myuser/bin/startvlc.sh) with this code:
#/bin/bash
declare +x LD_LIBRARY_PATH
/usr/bin/vlc "$1"
Remember make it executable: chmod u+x /home/myuser/bin/startvlc.sh
This script "delete" the LD_LIBRARY_PATH variable that was set in the command line to launch amule:
LD_LIBRARY_PATH=/home/myuser/usr/local/lib/ LANG=en_US.UTF-8 ./amule --even-if-lfroen-complains-this-will-stay.
VLC will find the libraries that are installed in the system and not in your home directory
I hope this will be useful for you.
-
Hi all,
I was hoping this memleak bug was fixed in wxGTK v2.8 (if indeed it is there and not amule monolithic) but I just got it by it again after an uptime of 24 hours. Here is the top output I managed to capture before killing amule:
20:05:55 up 1 day, 27 min, 6 users, load average: 22.20, 29.16, 32.42
Tasks: 209 total, 14 running, 195 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.2%us, 99.1%sy, 0.0%ni, 0.0%id, 0.1%wa, 0.1%hi, 0.6%si, 0.0%st
Mem: 2074936k total, 2025640k used, 49296k free, 796k buffers
Swap: 1020116k total, 1020116k used, 0k free, 4948k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3743 xxxxx 20 0 2493m 1.8g 440 R 17 89.6 216:59.54 amule
0164 xxxxx 20 0 200m 1080 444 S 10 0.1 63:20.15 audacious
221 root 15 -5 0 0 0 D 12 0.0 10:06.68 kswapd0
5317 xxxxx 20 0 110m 5676 996 D 7 0.3 40:17.46 k3b
3291 root 20 0 110m 35m 232 R 6 1.7 73:50.54 X
10164 xxxxx 20 0 200m 1124 488 S 10 0.1 63:22.28 audacious
3648 xxxxx 20 0 56948 1528 400 D 6 0.1 4:00.23 gnome-netstatus
Here, FWIW, is the cas output:
aMule CVS Fri Nov 9 07:01:54 CET 2007 has been running for 23:03 h
xxxxx is connected to ### xxxxxx ### [xxx.xxx.xxx.xxx] with HighID | Kad: ok
Total Download: 2.08 TB, Upload: 3.90 TB
Session Download: 1.99 GB, Upload: 1.10 GB
Download: 37.4 kB/s, Upload: 21.2 kB/s
Sharing: 853 file(s), Clients on queue: 925
Time: Nov 26 2007, 20:23
As I have indicated before, this bug does not occur with amuled.
$ rpm -qa | grep wx
wxGTK-2.8.4-6.fc8
wxGTK-devel-2.8.4-6.fc8
wxGTK-gl-2.8.4-6.fc8
compat-wxGTK26-2.6.3-8
$ uname -r
2.6.23.1-49.fc8[code]
[/code]
-
Wow, this mem leak is still in the code; it bit me in the ass again this evening:
top - 21:27:23 up 16 days, 8:05, 11 users, load average: 31.47, 33.96, 37.49
Tasks: 341 total, 2 running, 338 sleeping, 0 stopped, 1 zombie
Cpu(s): 0.2%us, 0.9%sy, 0.0%ni, 19.8%id, 79.0%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 3355684k total, 3265356k used, 90328k free, 1856k buffers
Swap: 2048248k total, 2048248k used, 0k free, 13428k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5621 xxxx 20 0 2728m 2.2g 552 D 0.0 67.6 1675:46 amule
5459 xxxx 20 0 678m 129m 144 S 0.0 3.9 433:35.64 java
18557 xxxx 20 0 713m 65m 412 S 0.0 2.0 47:15.33 thunderbird-bin
20162 xxxx 20 0 624m 56m 1204 D 0.6 1.7 32:25.90 chrome
28988 xxxx 20 0 524m 43m 260 S 0.0 1.3 46:33.79 acroread
2291 root 20 0 241m 37m 632 D 0.0 1.1 528:52.23 Xorg
15502 xxxx 20 0 146m 33m 264 S 0.0 1.0 0:12.53 chrome
...
As you can see, when I finally killed it, amule was using 67% of the 4G RAM.
$ rpm -qa | egrep "(amule|wx)" | sort
amule-2.2.6-1.fc12.i686
amule-nogui-2.2.6-1.fc12.i686
wxBase-2.8.10-7.fc12.i686
wxGTK-2.8.10-7.fc12.i686
wxGTK-devel-2.8.10-7.fc12.i686
wxGTK-gl-2.8.10-7.fc12.i686
wxGTK-media-2.8.10-7.fc12.i686
-
How fast is your upload speed?
If it's very fast (LAN speed) try the SVN version. There is a fix for such a memleak problem in it.