#13 Allocation of shared memory fails
I ran into a weird bug. I found that after about 1.5
Mpkts with the attached configuration capture dies
reporting an out of memory message. However the memory
usage is only 9MB out of 32MB. I suspect either a
memory leak or a problem with memory fragmentation.
I was running CoMo with one capture client executed
with curl "http://localhost:44444/services/trace" > /
dev/null
It seems to be reproducible and it seems that the
segmentation fault happens in lib/bitmap.c:110.
This is the shared memory map after the crash:
(gdb) p *shared_mem
$4 = {map = {{next = 0x0, size = 2048, count = 2},
{next = 0xb7d941a4,
size = 41504, count = 1}, {next = 0xb7d93df4,
size = 42448, count = 1}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0xb7d8414c, size = 1072, count = 1},
{next = 0x0, size = 48,
count = 0}, {next = 0x0, size = 56, count = 0},
{next = 0x0, size = 56,
count = 0}, {next = 0x0, size = 56, count = 0},
{next = 0x0, size = 56,
count = 0}, {next = 0x0, size = 56, count = 0},
{next = 0x0, size = 56,
count = 0}, {next = 0xb7d84fcc, size = 52, count
= 1}, {next = 0x0,
size = 56, count = 0}, {next = 0xb7d84c70, size
= 24, count = 1}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0x0, size = 56, count = 0}, {next = 0x0,
size = 56, count = 0}, {
next = 0xb7d7c940, size = 48, count = 1}, {next
= 0x0, size = 56,
count = 0}, {next = 0xb7daf7b4, size = 20, count
= 2}, {next = 0x0,
size = 56, count = 0}, {next = 0xb7d863f0, size
= 4096, count = 1}, {
next = 0xb7d7c4f8, size = 64, count = 1}, {next
= 0xb795d414, size = 64,
count = 1}, {next = 0xb7b440e0, size = 64, count
= 1}, {
next = 0xb75210bc, size = 64, count = 1}, {next
= 0xb6dedaa4, size = 64,
count = 1}, {next = 0x0, size = 0, count =
0} <repeats 2012 times>},
low = 0xb5e4d010 "", high = 0xb7e47000 "", usage =
10695404, peak = 10750724}
I don't have a core dump for this crash.
Config file