Skip to Content.
Sympa Menu

ndt-users - Re: Middleboxes: "protocol error!"

Subject: ndt-users list created

List archive

Re: Middleboxes: "protocol error!"


Chronological Thread 
  • From: Richard Carlson <>
  • To: ,
  • Subject: Re: Middleboxes: "protocol error!"
  • Date: Thu, 08 Mar 2007 11:19:26 -0500

Hi Claudio;

Are you running the web100srv process as root? The link type detection mechanism requires access to the raw network interface, and this is usually restricted to root. Send us the output from the "ps axuw | egrep "fakewww|web100srv" | grep -v egrep" command.

Where are you writing the web100srv.log file? I think there is a bug in the code that causes the "protocol error" message if the write to the web100srv.log file fails. I have an unreleased patch for this, and I'll try and get it out shortly. In the mean time, check the file system and make sure the log file exists and there is room in the file system.

The other thing you can do is run the web100srv process with some debugging options.

Simply run "sudo /usr/local/sbin/web100srv -d" and the debugging output will appear in your terminal window. Adding more '-d' flags (-ddd) will get you more details (3 levels with a -ddd). Start with 1 level and add more if nothing shows up.

Regards;
Rich Carlson
At 08:53 AM 3/8/2007,

wrote:
Hi,

I have some problems with NDT, but the main one reffer checking middleboxes. NDT work fine but sometimes during checking middleboxes web100clt (and also the java applet) return a "Protocol error!". Test was perform on localhost and also connecting by another pc and results it's the same. If I use web100clt with the '--disablemid' option, all (quite all...) works perfectly. NDT version is 3.3.19 on a Scientific Linux CERN 4.4 with a kernel 2.6.10 web100 patched. The userland library is 1.6 with the patch for IPv4 addresses mapped on IPv6. Follow up test results with '--disablemid' enabled:

Testing network path for configuration and performance problems -- Using IPv4 address
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client to server) . . . . . 982.12 Mb/s
running 10s inbound test (server to client) . . . . . . 424.24 Mb/s
Server unable to determine bottleneck link type.
Server 'xxxxxxxxxxxxxxx' is not behind a firewall. [Connection to the ephemeral port was successful]
Client is not behind a firewall. [Connection to the ephemeral port was successful]

------ Web100 Detailed Analysis ------

Web100 reports the Round trip time = 0.29 msec;the Packet size = 16384 Bytes; and
No packet loss - but packets arrived out-of-order 0.02% of the time.
This connection is receiver limited 99.68% of the time.

Web100 reports TCP negotiated the optional Performance Settings to:
RFC 2018 Selective Acknowledgment: ON
RFC 896 Nagle Algorithm: ON
RFC 3168 Explicit Congestion Notification: OFF
RFC 1323 Time Stamping: ON
RFC 1323 Window Scaling: ON; Scaling Factors - Server=2, Client=2
The theoretical network limit is 430403.47 Mbps
The NDT server has a 216 KByte buffer which limits the throughput to 5818.97 Mbps
Your PC/Workstation has a 67 KByte buffer which limits the throughput to 1792.41 Mbps
The network based flow control limits the throughput to 2155.17 Mbps

Client Data reports link is ' -1', Client Acks report link is ' -1'
Server Data reports link is ' -1', Server Acks report link is ' -1'
CurMSS: 16384
X_Rcvbuf: 221184
X_Sndbuf: 221184
AckPktsIn: 32393
AckPktsOut: 0
BytesRetrans: 0
CongAvoid: 0
CongestionOverCount: 0
CongestionSignals: 0
CountRTT: 32387
CurCwnd: 32768
CurRTO: 201
CurRwinRcvd: 17780
CurRwinSent: 32768
CurSsthresh: 2147483647
DSACKDups: 0
DataBytesIn: 0
DataBytesOut: 532361884
DataPktsIn: 0
DataPktsOut: 32392
DupAcksIn: 7
ECNEnabled: 0
FastRetran: 0
MaxCwnd: 81920
MaxMSS: 16384
MaxRTO: 211
MaxRTT: 40
MaxRwinRcvd: 68132
MaxRwinSent: 32768
MaxSsthresh: 0
MinMSS: 16364
MinRTO: 201
MinRTT: 0
MinRwinRcvd: 1072
MinRwinSent: 32768
NagleEnabled: 1
OtherReductions: 2
PktsIn: 32393
PktsOut: 32392
PktsRetrans: 0
RcvWinScale: 2
SACKEnabled: 3
SACKsRcvd: 0
SendStall: 0
SlowStart: 3
SampleRTT: 0
SmoothedRTT: 1
SndWinScale: 2
SndLimTimeRwin: 9996291
SndLimTimeCwnd: 561
SndLimTimeSender: 31951
SndLimTransRwin: 2
SndLimTransCwnd: 1
SndLimTransSender: 1
SndLimBytesRwin: 532230552
SndLimBytesCwnd: 82180
SndLimBytesSender: 49152
SubsequentTimeouts: 0
SumRTT: 9406
Timeouts: 0
TimestampsEnabled: 1
WinScaleRcvd: 2
WinScaleSent: 2
DupAcksOut: 0
StartTimeUsec: 789820
Duration: 10030003
c2sData: -1
c2sAck: -1
s2cData: -1
s2cAck: -1
half_duplex: 0
link: 100
congestion: 0
bad_cable: 0
mismatch: 0
spd: 424.67
bw: 430403.47
loss: 0.000001000
avgrtt: 0.29
waitsec: 0.00
timesec: 10.00
order: 0.0002
rwintime: 0.9968
sendtime: 0.0032
cwndtime: 0.0001
rwin: 0.5198
swin: 1.6875
cwin: 0.6250
rttsec: 0.000290
Sndbuf: 221184
aspd: 0.00000
CWND-Limited: 0.00

And now the second problem...
"Server unable to determine bottleneck link type." Searching in some previous articles in mailing list i found something about it (eth0 is the only network device and enter properly in promiscuous mode) but not work. Any idea?

Regards,

Claudio.
(i'm sorry for my english...)

------------------------------------



Richard A. Carlson e-mail:

Network Engineer phone: (734) 352-7043
Internet2 fax: (734) 913-4255
1000 Oakbrook Dr; Suite 300
Ann Arbor, MI 48104



Archive powered by MHonArc 2.6.16.

Top of Page