Add option to disablekeepalive

[sfunkhouser]

I ran into the issue that was noted in #35. When connecting locally from the US east coast to a switch in Singapore my connections would pile up. However, when connecting to a switch in New York, I did not run into the issue.

Using the debugging from the issue:

lsof -i -n -P |grep com.docke | awk '{ print $9 }' |awk -F'>' '{print $2}' | sort | uniq -c | sort -n

I saw every time I ran a command a new connection would be created, but the connection would stay around, building up connections, and goroutines. When timing these calls, the would take between 2-7 seconds:

server_1  | 2021/07/09 15:04:10 load: 2.3043784s
server_1  | 2021/07/09 15:04:10 load: 2.3117265s
server_1  | 2021/07/09 15:04:10 load: 2.3401141s
server_1  | 2021/07/09 15:04:10 load: 2.3514422s
server_1  | 2021/07/09 15:04:10 load: 3.1078603s
server_1  | 2021/07/09 15:04:11 load: 3.1816625s
server_1  | 2021/07/09 15:04:11 load: 3.1809949s
server_1  | 2021/07/09 15:04:11 load: 3.2186101s
server_1  | 2021/07/09 15:04:11 load: 3.3545421s
server_1  | 2021/07/09 15:04:11 load: 3.4516948s
server_1  | 2021/07/09 15:04:11 load: 3.4777806s
server_1  | 2021/07/09 15:04:11 load: 3.4740437s
server_1  | 2021/07/09 15:04:11 load: 3.6690569s
server_1  | 2021/07/09 15:04:11 load: 3.9385031s
server_1  | 2021/07/09 15:04:14 load: 6.6771321s
server_1  | 2021/07/09 15:04:14 load: 6.6883083s

When connecting to switches in NY, these were all under 1 second, and the connections would immediately clear up, and we had no pile up of goroutines

server_1  | 2021/07/09 15:58:19 load: 708.837ms
server_1  | 2021/07/09 15:58:19 load: 710.5202ms
server_1  | 2021/07/09 15:58:19 load: 715.5299ms
server_1  | 2021/07/09 15:58:19 load: 717.2996ms
server_1  | 2021/07/09 15:58:19 load: 722.4879ms
server_1  | 2021/07/09 15:58:19 load: 716.7412ms
server_1  | 2021/07/09 15:58:19 load: 764.4918ms
server_1  | 2021/07/09 15:58:19 load: 756.8071ms
server_1  | 2021/07/09 15:58:19 load: 770.9526ms
server_1  | 2021/07/09 15:58:19 load: 770.7404ms

Adding the DisableKeepAlive option cleared up the connections immediately after they were complete.

	node.GetConnection().SetDisableKeepAlive(true)

[cheynearista]

👍