aMule Forum
English => en_Bugs => Topic started by: monohouse on January 17, 2009, 03:10:39 AM
-
hardly a bug but this is still the best category for it, what could be the causes of leakage in amuled ?
ArchLinux 2.6.27.6 i686 Pentium III (Katmai)
VIRT/RES : aMuleD Uptime (not system)
------------------------------------------------------
/62.5 : 1 Days 09:36 hours
/115 : 2 Days 20:47 hours
/126 : 4 Days 20:31 hours
231/130 : 7 Days 13:08 hours
243/146 : 10 Days 08:50 hours
(http://manoa.flnet.org/Pics/amuled.png)
note: lines between points are not true, but should not matter because the points are correct
it would seem that most memory is consumed in the first 3 days at a rate which would seem to be between 1.7 and 1.8 MB/hour , after that it slows down to somewhere between 1.0 and 1.4 MB/hour RSS+VIRT and stays, I wonder if later it slows down more and gets to zero, if VIRT includes RSS then the calculation is incomplete but it still cannot be less than RSS
file buffer size: 1.5MB
queue size: 500
Number of Shared Files: 62
Total size of Shared Files: 67.52 GB
Average file size: 1.09 GB
Active Connections (estimate): 16
Max Connection Limit Reached: Never
Average Connections (estimate): 14.8372
Peak Connections (estimate): 61
I can provide additional statistics but I thought these were more important than others, if there is no idea, then I hope this information can prove useful to developers
-
Question: are you only seeing this leak in amuled? I am running amule 2.2.3 (the 'amule' program, not 'amuled') under FreeBSD 7.1-stable / amd64, on a machine with 3 GB physical RAM and a 5 GB swap partition. In less thn 24 hours, amule gets killed by the operating system because it is out of swap. Using 'top' as a tool, I notice that while SIZE doesn't change, RES (resident size) grows steadily hour by hour.
-
I wrote some patches for minor memory leaks and now I have my router running for 3 days. Considering that router have only 64mb no swap and 35mb used by cache (/proc/meminfo) I can say that I have so such problems. Can you try a recent snapshot ??