That sounds like something got corrupted in your .met files or something.
You can track down the source of those malloc messages by putting a breakpoint on malloc_printf and printing backtraces when it is hit. Then you can continue. When you feel you have a representative sampling of the source(s) of these messages, you can delete or disable the breakpoint. Here's a set of gdb commands which will automate the process of grabbing a few backtraces on malloc_printf for you:break malloc_printf
set $count = 0
commands
silent
if $count < 8
bt
set $count = $count + 1
end
cont
end[EDITED to include the all-important increment step.]
Note that even after the 8 backtraces, the breakpoint is still in effect, it's just continuing immediately. This means there's a small performance hit for each subsequent malloc message printed. If you like, just Ctrl-C to interrupt the program, disable the breakpoint, and continue. I would have had the breakpoint disable itself after the 8 backtraces, but gdb seems a bit buggy. When I've tried that in the past, it's crapped out on me.
After that, you might consider renaming your ~/.aMule directory and starting fresh. Start aMule once so it will create the necessary files in that directory and then you can move your Temp files back in to test.
