daylight
12-17-2002, 08:28 PM
I'm running the linux client v1.0.2 and have noticed a problem. When I type "netstat -pvat" I can see that the sb client is not completely closing the connections it makes to the seventeenorbust server. On my twin processor machine which has been running two copies of the client for five days, I've got 247 sockets still in use from this. On a single processor running one copy, I've got 115 half open sockets. Since the client only connects to the server infrequently, its only slowly increasing. But, sooner or later, it could actually use up enough resources to become a denial of sevice problem. Here are a couple lines from the output of "netstat -pvat | grep '/sb'":
tcp 1 0 rocket.digitaleve:45132 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:45127 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46712 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46699 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46686 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46656 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46660 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46645 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46589 sb.pns.net:1717 CLOSE_WAIT 29009/sb
This is most likely a simple one line bug caused by the client not closing the socket correctly. Although it affects both Linux machines I'm using, I suppose it also could just be something unique to my systems. Is anyone else seeing this?
On completely different issue, I think it would be interseting if the faq included information about how the back end system you are using is setup and what software/hardware you are using for the graphs, web server, database etc as well as the number of hits per day and how much load the server is under.
tcp 1 0 rocket.digitaleve:45132 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:45127 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46712 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46699 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46686 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46656 sb.pns.net:1717 CLOSE_WAIT 29009/sb
tcp 1 0 rocket.digitaleve:46660 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46645 sb.pns.net:1717 CLOSE_WAIT 29008/sb
tcp 1 0 rocket.digitaleve:46589 sb.pns.net:1717 CLOSE_WAIT 29009/sb
This is most likely a simple one line bug caused by the client not closing the socket correctly. Although it affects both Linux machines I'm using, I suppose it also could just be something unique to my systems. Is anyone else seeing this?
On completely different issue, I think it would be interseting if the faq included information about how the back end system you are using is setup and what software/hardware you are using for the graphs, web server, database etc as well as the number of hits per day and how much load the server is under.