Yessssss, it seems that the TCP race was my problem. :-)
I could crash 2.0.33 and some pre-2.0.34* with heavy network-load; now
I tortured one of my servers running *pre14 very very *VERY* hard
for ~1h and it stayed rock-solid! (>200 procs and up to 2K(!) TCP conns in
some state simultaneously on a 486DX4/16MB, sysload near 100.0). :-)
All previous versions I tried never survived that abuse for more
than 15 minutes. :-)
Great work, thanks!
pm 8-)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu