Ok, I think I found the problem. It appears to be at least partially self-inflicted. I was modifying the kernel buffer settings via /proc/* to improve performance. I noticed that when I used kernel-defaults, the TCP connection hangs stopped (and ran ~20% slower). My personal opinion is that this is still a bug somewhere in TCP, but at least I have a work-around.
Since these settings appear somewhat delicate, it might be nice if the kernel could warn about configurations that could cause TCP deadlocks and such. My old settings that exhibitted the problem were: my $netdev_max_backlog = 5000; my $wmem_max = 4096000; my $wmem_default = 128000; my $rmem_max = 8096000; my $rmem_default = 128000; my $tcp_rmem_min = 4096; my $tcp_rmem_default = 256000; my $tcp_rmem_max = 30000000; my $tcp_wmem_min = 4096; my $tcp_wmem_default = 256000; my $tcp_wmem_max = 30000000; my $tcp_mem_lo = 20000000; my $tcp_mem_pressure = 30000000; my $tcp_mem_high = 60000000; The new settings that appear to work around the TCP connection hang, and still give decent performance are below. Documentation being thin, I used the suggested values in Documentation/networking/ixgb.txt as hints: my $netdev_max_backlog = 30000; my $wmem_max = 524287; my $wmem_default = 128000; my $rmem_max = 524287; my $rmem_default = 128000; my $tcp_rmem_min = 4096; my $tcp_rmem_default = 256000; my $tcp_rmem_max = 5000000; my $tcp_wmem_min = 4096; my $tcp_wmem_default = 256000; my $tcp_wmem_max = 5000000; my $tcp_mem_lo = 10000000; my $tcp_mem_pressure = 20000000; my $tcp_mem_high = 30000000; Thanks, Ben -- Ben Greear <[EMAIL PROTECTED]> Candela Technologies Inc http://www.candelatech.com - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
