Hi all,
My COS 6.4 box reports a lot of bandwidth being used with an unknown origin (average of 2Mbps).
This interferes with my game servers that report 4 to 8 times higher pings then before.
The problem is not a constant ... sometimes after a few hours it disappears by itself.
Can someone please guide me, to how I could troubleshoot this ... ?!?
Greetings,
John.
Ps.
This is what it looks like when the problem disappears (and reappears) without making any changes:
My COS 6.4 box reports a lot of bandwidth being used with an unknown origin (average of 2Mbps).
This interferes with my game servers that report 4 to 8 times higher pings then before.
The problem is not a constant ... sometimes after a few hours it disappears by itself.
Can someone please guide me, to how I could troubleshoot this ... ?!?
Greetings,
John.
Ps.
This is what it looks like when the problem disappears (and reappears) without making any changes:
Share this post:
Responses (8)
-
Accepted Answer
Hi all,
I was wondering if someone could help me with this problem.
Indefinite shutting down the Web server on the standalone COS box is not acceptable.
I have the Intranet and several other services running on it.
(Dolphin community, Galery 3, Joomla ... etc)
Maybe if someone can guide me to the log files to see what is going on with port 80.
Please advice,
John
Ps. I am willing to grant temporary access to the box if this would help. -
Accepted Answer
Hi all,
I think I discovered the cause of the massive bandwidth use.
I didn't really understand the results of iptraf, but when I noticed the ip address of the stand alone COS 5.2 server (192.168.1.100), I tried something ...
I rebooted the stand alone COS box and suddenly the bandwidth problem disappeared.
After it booted up again the problem returned.
This time the problem started around midnight.
Now that I think I know where this erratic behavior originates, I wonder what I can do about it ... besides shutting it down.
Please advice.
John
EDIT:
After temporary shutting down the web server on the stand alone COS box, the problem disappeared, but that is not a solution I feel happy about. -
Accepted Answer
Hi Tim,
Sorry for the late response and thanks for your advice.
After I confirmed that all clients on the network where shut down last weekend, the problem did not return again.
I checked the program iptraf, but because the problem did not return I don't know what to look for there.
Currently the pings look normal, depending on the distance too the clients:
If the problem returns I will try to see what is causing it.
Greetings,
John
PS.
eth1 is my LAN interface and eth0 is connected to the Internet/router.
I have specified 15/2 Mbit down and up in the Bandwidth Manager (the maximum bandwidth according to the router settings). -
Accepted Answer
Odd!
Sorry John - I meant the iptraf command which gives a live view of traffic passing your interfaces. You should be able to identify where the traffic is originating from
eth1 is your LAN interface? as your graph suggests that eth0 is your WAN... have you specified any kind of bandwidth limits on this interface? -
Accepted Answer
Hi Tim,
I was uploading a few large files while some clients where surfing on the Internet with the following footprint:
As you can see eth1-recv is maxing out at approximately 2Mbps.
This did not result in high pings on my game servers, so that means that the problem I described is caused by something else.
Please advice,
John -
Accepted Answer
Hi Tim,
The problem has returned and the first part of the output of netstat is this:
# netstat
Active Internet connections (w/o servers)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 1 0 localhost:40565 localhost:ldap CLOSE_WAIT
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 localhost:ldap localhost:59953 ESTABLISHED
tcp 1 0 localhost:40739 localhost:ldap CLOSE_WAIT
tcp 0 0 system.domain.lan:ssh system.gamesangolin-laser ESTABLISHED
tcp 1 0 localhost:40504 localhost:ldap CLOSE_WAIT
tcp 0 0 192.168.178.21:81 <internet ip>:<port> ESTABLISHED
tcp 1 0 localhost:40625 localhost:ldap CLOSE_WAIT
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 system.domain.lan:squid system.gameserver.la:anet-b ESTABLISHED
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 system.domain.lan:squid system.gameserver.la:anet-l ESTABLISHED
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 system.domain.lan:ssh system.gameserver.la:dj-ilm ESTABLISHED
tcp 0 0 system.domain.lan:squid system.gameserver.la:anet-h ESTABLISHED
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 system.domain.lan:squid system.gameserver.la:webtie ESTABLISHED
tcp 0 0 192.168.178.21:81 <internet ip>:<port> TIME_WAIT
tcp 0 0 localhost:59953 localhost:ldap ESTABLISHED
tcp 1 0 localhost:38533 localhost:ldap CLOSE_WAIT
tcp 0 0 ::ffff:192.168.178.21:40159 server1611.teamviewer.:http ESTABLISHED
tcp 0 0 ::ffff:192.168.178.21:52128 server10006.teamviewer:http ESTABLISHED
tcp 0 0 ::ffff:192.168.178.21:40158 server1611.teamviewer.:http ESTABLISHED
tcp 0 0 ::ffff:192.168.178.21:52126 server10006.teamviewer:http ESTABLISHED
udp 0 0 192.168.178.21:49088 system.modem.lan:nat-pmp ESTABLISHED
Active UNIX domain sockets (w/o servers)
There is some more info following after this, but it looks like that is not important.
Please advice,
John -
Accepted Answer
Hi Tim,
Thanks for responding.
When I manually update one of the clients the problem does not occur, so I doubt if this is the cause.
Even when I download a big file, the signature on the Bandwidth Viewer looks totally different. (average 10Mbps)
The bandwidth speeds are 10Mbps down and 1Mbps up. (according to the router settings there is a 20/2 Mbps down/up connection)
I attempted to get some info with the Network Visualiser app, but the only sources I currently see is of the ip address of eth0 (192.168.178.21 / external / DHCP).
I am new to the netstat command line tool and I must admit that the information it provides does not help me, but the next time the problem occurs I will post its results.
Greetings,
John -
Accepted Answer
Hi John, perhaps automatic updates somewhere on your network? usually the saturation of upload bandwidth on asymetric connections causes increased latency (as the packets have to queue up at the modem). What bandwidth speeds do you get up/down? Use the network visualiser app or the netstat command line tool to identify which host is causing this traffic?
Please login to post a reply
You will need to be logged in to be able to post a reply. Login using the form on the right or register an account if you are new here.
Register Here »