ndt-project / ndt Goto Github PK

Original issue 17 created by aaronmbr on 2010-02-27T01:40:01.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in both server and client.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 65536 65536" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 8912 8912" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. Check the statistics window.

What is the expected output? What do you see instead?
Since the send buffer is low in both server and client which are limiting
the throughput both should be reported but only server send buffer is
reported in the
statistics window, client send buffer problem not reported.

Please provide any additional information below.

Original issue 48 created by aaronmbr on 2010-06-10T22:35:53.000Z:

What steps will reproduce the problem?
This problem occurred when tested with multiple simultaneous clients.

1. Start test with 70 or 80 or 100 simultaneous clients.
2. Check the messages displayed by clients.
3. Run this test for many times.
4. After some time tests are getting failed. Middle box test port received by client shown as 0 and test failed.

What is the expected output? What do you see instead?
When 70/80/100 Clients start testing simultaneously all the clients should
complete the test successfully or should get Server Busy.

Tests are getting failed after client receiving the middlebox test port as 0. This problem occurred 2 times during our testing.
At the time problem occurred log files are captured and attached.

What version of the product are you using? On what operating system?
NDT server 3.6.3
Client : Linux (Ubuntu 9)

Original issue 8 created by aaronmbr on 2010-02-27T01:27:14.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in both server and client.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. Check the statistics window.

What is the expected output? What do you see instead?

It is observed that in statistics window "Information: Network Middlebox is
modifying MSS variable" message displayed. In this setup the MSS is not
modified by middlebox (switch) but the report says MSS is modified by the
middlebox. MSS value is 1488 in this case. Since the TCP option timestamps,
window scaling are enabled MSS is 1488.
If the TCP options timestamps and window scaling are disabled the MSS value
is displayed as 1460 and the messaged "Network Middlebox is modifying MSS
variable" is not displayed.

Please provide any additional information below.

Original issue 44 created by aaronmbr on 2010-04-19T21:57:41.000Z:

What steps will reproduce the problem?

1.Run the Web100 server.
2.Run the script to log Memory usage in the server.
2.Set the Bandwidth to 100mbit (10 clients X 10 Mbit) in 2 clietns systems.
3.Run the script to invoke NDT clients with parameters 10 number of clients
and the server address in both client systems to simulate 20 clietns.
4.Monitor the clients execution
5.After completion of the test check the log files created to verify that
all clients are completed the test successfully.
6.Check the CPU usage

What is the expected output? What do you see instead?
CPU usage during Middlebox testing is some times more (70% to 100%). It is
observed peak CPU usage during middlebox testing during performance
testing. This is not occuring everytime.
During other tests of the server CPU usage is not more (less then 25% for
the load 20 clients each having 25mbits bandwith).

What version of the product are you using? On what operating system?
NDT server 3.6.1
Clients : Linux (Ubuntu 9)

Please provide any additional information below.

CPU usage during test run with 20 clients with 10mbps bandwidth to each client

Time CPU usage (%) Test Stage
1 0 Stat
2 1
3 0
4 89 Middle Box Testing
5 94
6 100
7 100
8 98
9 91
10 91
11 1 Firewall Testing
12 0
13 4 C2S Testing
14 5
15 3
16 9
17 10
18 10
19 12
20 13
21 9
22 11
23 13
24 15
25 22
26 10 S2C Testing
27 4
28 1
29 17
30 28
31 2
32 2
33 11
34 11
35 17
36 18
37 2
38 0
39 0
40 4

Original issue 43 created by aaronmbr on 2010-03-12T01:32:12.000Z:

A user saw this java script error on the NDT site directly (On IE). Getting same problem on fcc/m-
lab test.

Original issue 15 created by aaronmbr on 2010-02-27T01:37:15.000Z:

What steps will reproduce the problem?
1.Open the browser
2.Click “Start” button in the applet.
3.Click “Option” button

What is the expected output? What do you see instead?
Expected Output:
Option button pop up window should be displayed.

Observed Output:
1."Report Problem" button is displaying in disable mode after first time
clicking the "Report Problem" button in Applet. After clicking the “Report
Problem” tab focus is moving to “Option” button.

1. After opening the "Statistics" , "more details" pop up windows, tab
   order focus is not working properly. After closing the "Statistics" pop up
   window, focus is moving to "More details" window instead of displaying in
   "Statistics" window.

Original issue 32 created by aaronmbr on 2010-03-05T23:46:49.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Click the start button to start the test and wait till client complete
   the test.
4. Check the Buttons on the applet.

What is the expected output? What do you see instead?

All buttons should be displayed in enabled condition but the display shows
them as disabled. But when clicked on them they are working.

What version of the product are you using? On what operating system?
NDT server 3.6.1
Client : MAC (10.5)
Browser: Firefox (3.0)

Please provide any additional information below.

Original issue 30 created by aaronmbr on 2010-03-05T18:26:50.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Down the network interface eth0.
4. Set autoneg option of eth0 to off with ethtool.
5. Up the network interface eth0.
6. Set the network interface duplex mode as off and speed as 100Mbps with
   ethtool.
7. Restart the network subsystem /etc/init.d/network restart.
8. Open the browser and open the NDT home page.
9. Start test by clicking Start button and check the results on main window
   and statistics window.

What is the expected output? What do you see instead?
In Statistics window duplex mode should be displayed as half-duplex and
report duplex mismatch since server is in full duplex mode.
But duplex mode is displayed as full duplex and not reported duplex mismatch.

What version of the product are you using? On what operating system?
NDT server 3.6.1
Client : Linux (Fedora 8)

Please provide any additional information below.

Original issue 38 created by aaronmbr on 2010-03-10T03:25:44.000Z:

Write a simple init script instead of relaying on remote commands to control the start and stop of
donar.

Original issue 45 created by aaronmbr on 2010-06-05T20:40:07.000Z:

What steps will reproduce the problem?

1. Run server on machine with wlan interface.
2. Start test on another machine. Traffic must go via server's wlan interface.
3. Look at results on the client machine.

What is the expected output? What do you see instead?

EXPECTED:

Test finishes successfully and the results are printed correctly.

CURRENT:

• client machine:

Protocol error - expected results! got ' 1', msgtype=2

• server machine (console output with multiple -d options):

<------------ 2429 ------------->
Finished testing C2S = 17.10 Mbps, S2C = 16.46 Mbps
Signal 11 received by process 2429
DEBUG, caught SIGSEGV signal and terminated process (2429)

Original issue 31 created by aaronmbr on 2010-03-05T20:40:10.000Z:

What steps will reproduce the problem?
1.Set the ECN option field as “ON” in client using the below command.
sudo echo "1" > /proc/sys/net/ipv4/tcp_ecn
2.check the ECN option field is setted to “ON” or not using the below command.
sudo cat /proc/sys/net/ipv4/tcp_ecn
3.Set the ECN option field as “ON” in server using the below command.
sudo echo "1" > /proc/sys/net/ipv4/tcp_ecn
4.check the ECN option field is setted to “ON” or not in server using the
below command.
sudo cat /proc/sys/net/ipv4/tcp_ecn
5.Open the browser with server address.
6.start the test by clicking the “Start” button
7.after completion of the test, open the statisticks window and check the
status of “Explicit Congestion Notification” information.

What is the expected output? What do you see instead?
Expected Output:
Explicit Congestion Notification option should be display as “ON”.

Observed Output:
Explicit Congestion Notification option always displaying “OFF”.

We have observed in tcpdump that ECT(1) (ECN Capable Transport(1) is setted
in IP header and ECE flag - ECN-Echo flag & CWR flag setted in tcp packet.
TCP dump is attached.

Original issue 39 created by aaronmbr on 2010-03-10T20:29:59.000Z:

Blanket report covering issue 12, issue 13, issue 14, issue 15, issue 32, and issue 33. Synopsis:

The applet:

• may not display properly (different platforms)
• May give the allusion of not loading/working properly
• may not function properly (information is not accurate, button work when they shouldn't)

See individual bug reports for details

Original issue 13 created by aaronmbr on 2010-02-27T01:36:03.000Z:

What steps will reproduce the problem?
1.Open the browser
2.Click “Start” button in the applet.
3.Click “Option” button

What is the expected output? What do you see instead?
Expected Output:
“options” pop up window should be displayed and able to selec the options
in the pop up window.

Observed Output:

1. “Default Tests” option in Option button is always displaying in Disable
   mode. Not able to modify this parameter.

2.After executing the 1st time of the test(Click Start button), “prefer
Ipv6” is always displaying in Disable mode. We are able to modify this
option before starting the 1st time of the test only.

3.After selecting the “prefer Ipv6” option, Applet is displaying “Connected
to: 64.9.225.187-- Using IPv4 address”.

Browser : Forefox(Linux,MAC 7 Windows),Chrome(Linux & Windows)

NDT Version : 3.6.0

Original issue 9 created by aaronmbr on 2010-02-27T01:27:57.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the MTU value of the network interface to 512.
4. Click the start button to start the test and wait till client complete
   the test.
5. Check the statistics window.

What is the expected output? What do you see instead?

Throughput is less since the MTU value of the NIC changed to 512 which is
limiting the throughput but it is not pointed in the statistics window.

Please provide any additional information below.

Original issue 47 created by aaronmbr on 2010-06-10T22:30:31.000Z:

What steps will reproduce the problem?
This problem occured when tested with 70/80/100 simultanious clients.

1. Start test with 70 or 80 or 100 simultabnious clients from 2 client machines
2. Check the messages displayed by clients.
3. Run this test for few times.

What is the expected output? What do you see instead?
When 70/80/100 Clients start testing simultaniously all the clients should
complete the test successfully or get Server Busy.
But some clients are getting ServerFault message. This occured when the client received wait flag as 2 intially then received wait flag 1 after
that it received wait flag 9988 insted of 0.

What version of the product are you using? On what operating system?
NDT server 3.6.3
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 35 created by aaronmbr on 2010-03-06T00:04:02.000Z:

What steps will reproduce the problem?

1.Open the Firefox browser with MLAB server IP address.
2.Start the test by clicking the “Start” button.
3.Open any one of word document or PDF document or linux terminal and
minimize it.
3.open any one of the pop up window by selecting the “Statistics” or “More
Details” or “Option”.
4.Close the pop up window.

What is the expected output? What do you see instead?
Expected Output:
Pop up window should be closed and focus should move to the selected button.

Observed Output:
Pop up window is closing and one of the minimized dialog (document or linux
teminal) or backend window is displaying automatically.

Browser : Firefox , Safari
OS : MAC

Original issue 18 created by aaronmbr on 2010-02-27T01:45:33.000Z:

What steps will reproduce the problem?
1.pass the maximum number of clients runtime parameter to server using “x”
option.

What is the expected output? What do you see instead?
Expected Output:
Server should accept the runtime parameter “x” and initilize the server
with received runtime parameter.

Observed Output:

server is not accepting “x” runtime parameter and giving the below
mentioned error.

srinivasu@gsolorzano:/usr/src/NDT-3.6.0/ndt-3.6.0/src$ clear

srinivasu@gsolorzano:/usr/src/NDT-3.6.0/ndt-3.6.0/src$ ./web100srv -4 -x 20

./web100srv: invalid option -- x

Try `./web100srv --help' for more information.

srinivasu@gsolorzano:/usr/src/NDT-3.6.0/ndt-3.6.0/src$

srinivasu@gsolorzano:/usr/src/NDT-3.6.0/ndt-3.6.0/src$

Original issue 36 created by aaronmbr on 2010-03-09T03:30:04.000Z:

Currently NDT is using zlib to handle compression, specifically
trunk/src/logging.c makes calls to the zlib deflate() function to compress
the log files. This example code from an on-line tutorial, but this does
not make them 'gzip' files.

The http://www.zlib.net/zlib_how.html page contains the example used to
write the deflate routine. Pulling the zpipe.c file off of this web page,
and compiling it will be enough to uncompress the snaplog and tcpdump log
files.

The 'fix' in this case should be to note this, and change the extension to
note that these are not zip files.

Original issue 20 created by aaronmbr on 2010-02-27T02:10:07.000Z:

What steps will reproduce the problem?
1.Start test with 100 clients simultaneously connect to server
2.check the messages displayed by 100 clients

What is the expected output? What do you see instead?
Expected Output:
All 100 clients should be complete the Middlebox,Firewall , Outbound and
InBound tests

Observed Output:
Some times all clients are not completing the tests. Some clients waiting
indefinite time after receiving the wait flag.

NDT Version : 3.6.0
OS : Ubuntu

Original issue 24 created by aaronmbr on 2010-03-03T22:12:35.000Z:

What steps will reproduce the problem?
1.Start test with 100 clients simulataneously connect to server
2.check the messages displayed by 100 clients
3.Check the server status using “ps -ef | grep web100srv” command.

What is the expected output? What do you see instead?
Expected Output:
All 100 clients should be complete the Middlebox,Firewall , Outbound and
InBound tests.

Observed Output:

web100srv creating defunct process and after creating defunct process,
web100srv is not responding to clients even web100srv is in running mode.
At this movement, after restarting the web100srv also web100srv is not
responding to the client. Some times web100srv responding to clients but
very slowly responding.But after restarting the system , web100srv is
responding to the clients

we have observed this issue in our local MLAB server and more frequently.
Logs attached.

Original issue 11 created by aaronmbr on 2010-02-27T01:33:10.000Z:

What steps will reproduce the problem?
1.Start test with more than 100 clients simulataneously connect to server
2.check the messages displayed by clients

What is the expected output? What do you see instead?
Expected Output:
All clients should be complete the Middlebox,Firewall , Outbound and
InBound tests

Observed Output:
when multiple clients(More than 100 Clients) connecting to the server and
observed few of the clients displaying the message "Server process not
running: start web100srv process on remote server (64.9.225.187:3001)".

we have observed,server is in running mode when client receives the the
above mentioned messsage and accepting the new connections to other clients.

NDT Version : 3.6.0
OS : Ubuntu

Original issue 28 created by aaronmbr on 2010-03-05T18:22:47.000Z:

WinXP + FF

1. After fresh restart of FF,
   http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/
2. Denied permissions (hit cancel on security warning)
3. Clicked START

Result: Long delay during download portion resulting in a failure to
complete test.

NOTE: the same steps on Ubuntu + FF resulted in the test completing
successfully and normally

Expected: Not sure, but if there is a dependency on security permissions,
and a required permission was not granted, then failure should be graceful.
Why behavior is inconsistent across platforms should be understood.

TCP/Web100 Network Diagnostic Tool v3.6.1
click START to begin

** Starting test 1 of 1 **
Connecting to 'ndt.iupui.mlab3.lga02.measurement-lab.org'
[ndt.iupui.mlab3.lga02.measurement-lab.org/38.106.70.173] to run test
Connected to: ndt.iupui.mlab3.lga02.measurement-lab.org -- Using IPv4 address
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 19.53Mb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . Server Failed
while receiving data
S2C throughput test FAILED!
Protocol error! Expected test results, got: ###instead

click START to re-test

Java Console + "s" "t" "v" below
Java Plug-in 1.6.0_10
Using JRE version 1.6.0_10 Java HotSpot(TM) Client VM
User home directory = C:\Documents and Settings\topolskir

c: clear console window
f: finalize objects on finalization queue
g: garbage collect
h: display this help message
l: dump classloader list
m: print memory usage
o: trigger logging
q: hide console
r: reload policy configuration
s: dump system and deployment properties
t: dump thread list
v: dump thread stack
x: clear classloader cache

0-5: set trace level to <n>

security: property package.access value sun.
security: property package.access new value sun.,com.sun.javaws
security: property package.access value sun.,com.sun.javaws
security: property package.access new value sun.,com.sun.javaws,com.sun.deploy
security: property package.definition value null
security: property package.definition new value com.sun.javaws
security: property package.definition value com.sun.javaws
security: property package.definition new value com.sun.javaws,com.sun.deploy
security: property package.definition value com.sun.javaws,com.sun.deploy
security: property package.definition new value
com.sun.javaws,com.sun.deploy,com.sun.jnlp
security: property package.access value sun.,com.sun.javaws,com.sun.deploy
security: property package.access new value
sun.,com.sun.javaws,com.sun.deploy,org.mozilla.jss
security: property package.definition value
com.sun.javaws,com.sun.deploy,com.sun.jnlp
security: property package.definition new value
com.sun.javaws,com.sun.deploy,com.sun.jnlp,org.mozilla.jss
basic: Added progress listener:
sun.plugin.util.GrayBoxPainter$GrayBoxProgressListener@1e4457d
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Connecting
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar with
proxy=DIRECT
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/
with proxy=DIRECT
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Connecting
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar with
proxy=DIRECT
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/
with proxy=DIRECT
network: Downloading resource:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar
Content-Length: -1
Content-Encoding: null
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
security: Accessing keys and certificate in Mozilla user profile: null
security: Loading Root CA certificates from C:\Program
Files\Java\jre6\lib\security\cacerts
security: Loaded Root CA certificates from C:\Program
Files\Java\jre6\lib\security\cacerts
security: Loading Deployment certificates from C:\Documents and
Settings\topolskir\Application Data\Sun\Java\Deployment\security\trusted.certs
security: Loaded Deployment certificates from C:\Documents and
Settings\topolskir\Application Data\Sun\Java\Deployment\security\trusted.certs
security: Loading certificates from Deployment session certificate store
security: Loaded certificates from Deployment session certificate store
security: Validate the certificate chain using CertPath API
security: Obtain certificate collection in Root CA certificate store
security: Obtain certificate collection in Root CA certificate store
security: No timestamping info available
security: Found jurisdiction list file
security: Start checking trusted extension for this certificate
security: Start comparing to jurisdiction list with this certificate
security: The CRL support is disabled
security: The OCSP support is disabled
security: Checking if certificate is in Deployment denied certificate store
security: Checking if certificate is in Deployment permanent certificate store
security: Checking if certificate is in Deployment session certificate store
security: User has denied the priviledges to the code
security: Adding certificate in Deployment denied certificate store
security: Added certificate in Deployment denied certificate store
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Connecting
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar with
proxy=DIRECT
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/
with proxy=DIRECT
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
network: Cache entry not found [url:
http://ndt.iupui.mlab3.lga02.measurement-lab.org:7123/Tcpbw100.jar,
version: null]
basic: Applet loaded.
basic: Applet resized and added to parent container
basic: PERF: AppletExecutionRunnable - applet.init() BEGIN ; jvmLaunch dt
117118 us, pluginInit dt 9565907 us, TotalTime: 9683025 us
basic: Applet initialized
basic: Loading Java Applet ...
basic: Removed progress listener:
sun.plugin.util.GrayBoxPainter$GrayBoxProgressListener@1e4457d
basic: Applet made visible
basic: Starting applet
basic: Applet started
basic: Told clients applet is started
Couldn't set system property. Check your security settings.
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:3001/
with proxy=DIRECT
wait flag received = 0
Server version: 3.6.1
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:47962/
with proxy=DIRECT
1525888 bytes 1009.3520754092939 kb/s 12.094 secs
Sending '1009.3520754092938' back to server
calling in2Socket.getLocalAddress()
SFW: port=36822
SFW: testTime=1
Asking security manager for listen permissions...
SFW: oport=4452
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:36822/
with proxy=DIRECT
java.net.SocketTimeoutException: Accept timed out
at java.net.PlainSocketImpl.socketAccept(Native Method)
at java.net.PlainSocketImpl.accept(Unknown Source)
at java.net.ServerSocket.implAccept(Unknown Source)
at java.net.ServerSocket.accept(Unknown Source)
at Tcpbw100$OsfwWorker.run(Tcpbw100.java:1642)
at java.lang.Thread.run(Unknown Source)
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:45581/
with proxy=DIRECT
Send buffer size =8192
java.net.SocketException: Software caused connection abort: socket write error
170.06584418751524 kb/s outbound
network: Connecting http://ndt.iupui.mlab3.lga02.measurement-lab.org:39010/
with proxy=DIRECT
Couldn't get 3rd connection to: ndt.iupui.mlab3.lga02.measurement-lab.org

Original issue 1 created by aaronmbr on 2010-02-27T01:16:10.000Z:

What steps will reproduce the problem?
This problem occured when tested with 100 simultanious clients.

1. Start test with 100 simultabnious clients
2. Check the messages displayed by clients.
3. Some clients display protoclo error and exit at different stages (middle
   box testing, inbound/outbound testing).

What is the expected output? What do you see instead?
When 100 Clients start testing simultaniously all the clients should
complete the test successfully.
But some clients are exited in between the test dislying "Protocol Error"

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 21 created by aaronmbr on 2010-03-01T13:49:49.000Z:

Examine scripts that:

• Restart NDT to ensure it is running
• Rotate and rsync NDT logs off of the measurement hosts
• Delete old logs to prevent disks from filling on measurement hosts

Examine distribution methods:

• Improve system to move released code from development to production hosts

Original issue 4 created by aaronmbr on 2010-02-27T01:22:08.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the receive buffer size in the both server and client to 32K and
   trun off receive buffer auto tuning.
4. Click the start button to start the test and wait till client complete
   the test.
5. In Moredetails window receive buffer size of the server is displayed
   correctly as 46k and client recive buffer size displayed as 22K.

What is the expected output? What do you see instead?
Moredetails window receive buffer size of server and client should be
displayed as 32k but server receive buffer suize is displayed correctly as
46k but client recive buffer size displayed as 22K.

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 29 created by aaronmbr on 2010-03-05T18:25:04.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the client TCP parameters as bellow
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 131071 > /proc/sys/net/core/wmem_max
   echo 131071 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Set the server TCP parameters as bellow.
   echo 0 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 131071 > /proc/sys/net/core/wmem_max
   echo 131071 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
5. Open the browser and open the NDT home page.
6. Start test by clicking Start button and check the results on main window
   and in more details window.

What is the expected output? What do you see instead?
Normally the throughput detected should be less than or equal to the
throughput calculated in the more details window.
Server receive buffer limiting the throughput is less where as the outbound
throughput displayed in main window is more.

What version of the product are you using? On what operating system?
NDT server 3.6.1
Client : Linux (Fedora 8)

Please provide any additional information below.

Original issue 22 created by aaronmbr on 2010-03-02T20:57:46.000Z:

What steps will reproduce the problem?

This problem occurred when tested with 100 simultaneous clients.

1. Start test with 100 simultaneous clients
2. Check the messages displayed by clients.
3. After some time clients are not getting response. After 3 minutes
   client exited with error.

What is the expected output? What do you see instead?
Clients should continue test successfully or should get wait flag. But
after connected to the server clients are not getting response. After this
condition even we run single client it is not getting response. This bug
occurred few times.

What version of the product are you using? On what operating system?
NDT server 3.6.0 / 3.6.1
Client : Linux (Ubuntu 9)

Please use labels and text to provide additional information.
We took the tcpdump during this condition and attached.
After restarting the server only clients able to complete test with server.

Original issue 7 created by aaronmbr on 2010-02-27T01:26:11.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch

2. Start both servers (fakewww and web100)

3. Set the following tcp paramters in both server and client.
   Server
   echo 0 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 32768 32768" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "1" > /proc/sys/net/ipv4/tcp_ecn

   Client
   echo 0 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "0" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "1" > /proc/sys/net/ipv4/tcp_ecn

4. Click the start button to start the test and wait till client complete
   the test.

5. Check the statistics window.

What is the expected output? What do you see instead?
When promting the send buffer size problem current size of the send buffer
should be displayed.
But statistics window is displaying the send buffer size as 16K with is not
correct. In the more details page NDT server send buffer size X_Sendbuf
values is displayed correctly 32768 (32K)

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 42 created by aaronmbr on 2010-03-12T01:27:53.000Z:

When upgrading from 3.5.14 to 3.6.1 we noticed that the "-m" mode for allowing multiple
parallel clients stopped working.
With that option enabled, the inbound test fails as follows:

TCP/Web100 Network Diagnostic Tool v3.6.1
click START to begin

** Starting test 1 of 1 **
Connecting to 'lsmp2' [lsmp2/130.59.35.42] to run test
Connected to: lsmp2 -- Using IPv4 address
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 91.97Mb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . Protocol error! Expected test data, got:

ndt-3.6.1-debug-output.txt

TCP/Web100 Network Diagnostic Tool v3.6.1
click START to begin

** Starting test 1 of 1 **
Connecting to 'lsmp2' [lsmp2/130.59.35.42] to run test
Connected to: lsmp2 -- Using IPv4 address
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 91.97Mb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . Protocol error! Expected test data, got:
instead
S2C throughput test FAILED!
Protocol error! Expected test results, got: instead

click START to re-test

The debug log on the server:

: 1root@lsmp2[welti]; /usr/local/sbin/web100srv -d -a -m --ipv4
ANL/Internet2 NDT ver 3.6.1
Variables file = /usr/local/ndt/web100_variables
log file = /usr/local/ndt/web100srv.log
Admin file = /usr/local/ndt/admin.html
Debug level set to 1
server ready on port 3001
web100_init() read 69 variables from file
Initial counter Values Totalcnt = 0, Total Mismatch = 0, Total Bad Cables = 0
Updated counter values Totalcnt = 1, Total Mismatch = 0, Total Bad Cables = 0
...
Initial counter Values Totalcnt = 594, Total Mismatch = 0, Total Bad Cables = 0
Updated counter values Totalcnt = 595, Total Mismatch = 0, Total Bad Cables = 0
Individual counts = [33, 2, 4, 421, 30, 18, 41, 42, 4, 0, 0, 0, 0, 0, 0, 0]
Signal 17 received by process 9888
successfully locked '/tmp/view.string' for updating
sending '988488,0,989562,0,595,0,0,33,2,4,421,30,18,41,42,4,0,0,0,0,0,0,0,Mar 9
13:46:21,Jan 26 10:44:53' to tmp file
Starting test suite:
> Middlebox test
> Simple firewall test
> C2S throughput test
> S2C throughput test
<-- Middlebox test -->
WARNING: ephemeral port number was bound
-- port: 54794
Sending 1444 Byte packets over the network
Signal 17 received by process 9891
<-------------------->
<-- Simple firewall test -->
-- port: 43250
-- time: 1
-- oport: 53549
Unable to create connect socket.
<-------------------------->
<-- C2S throughput test -->
WARNING: ephemeral port number was bound
-- port: 46533
listening for Inet connection on testOptions->c2ssockfd, fd=3
Sending 'GO' signal, to tell client to head for the next test
Opening network interface 'eth0' for packet-pair timing
installing pkt filter for 'host maccw.switch.ch and port 53554'
Initial pkt src data = 806d804
New IPv4 packet trace started -- initializing counters
91973 kbps outbound
Signal USR1(10) sent to child [9897]
Signal 10 received by process 9897
03:19:30.019186 57 bytes read ' 1 1 0 39 205 78854 278 2 127 11 83 2 266.12 0 0 0 1 0 7'
from C2S monitor pipe
03:19:30.019186 65 bytes read ' 0 0 1 9 118 38235 112 0 1 58 3372 0 100.05 99 67 41740 0
3372 7' from C2S monitor pipe
<------------------------->
<-- S2C throughput test -->
Signal 17 received by process 9891
WARNING: ephemeral port number was bound
-- port: 50564
waiting for data on testOptions->s2csockfd
Opening network interface 'eth0' for packet-pair timing
installing pkt filter for 'host maccw.switch.ch and port 53569'
Initial pkt src data = 806d804
Signal 17 received by process 9891
New IPv4 packet trace started -- initializing counters
Signal 15 received by process 9891
Signal 17 received by process 9888
Signal 14 received by process 9899

Original issue 6 created by aaronmbr on 2010-02-27T01:24:51.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in both server and client.
   echo 0 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 32768 32768" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 16384" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "0" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "1" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. Check the staticstics window

What is the expected output? What do you see instead?
Since Throughput can be improved by increasing the recive and send buffers
messege should be displayed in statistics window.
But No message has been displayed syaing that by increasing the recive and
send buffers throughput can be increased. And also the buffer sizes
displayed in the more details page is also not correct. Server receive
buffer showed as 8K instead of 32k and client receive buffer is showed as
22k instead of 32k.

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 34 created by aaronmbr on 2010-03-06T00:01:33.000Z:

What steps will reproduce the problem?

1.Open the browser with MLAB server IP address.
2.start the test by clicking the “Start” button.
3.open the More details page and check the DataPktsIn and DataBytesIn fields.

What is the expected output? What do you see instead?
Expected Output:
Respective count should display of the DataPktsIn and DataBytesIn.

Observed Output:
DataPktsIn and DataBytesIn fileds are displaying as “0”.

Original issue 46 created by aaronmbr on 2010-06-08T13:22:09.000Z:

What steps will reproduce the problem?

1. Run an NDT test from a machine known to run NAT

What is the expected output? What do you see instead?
API function get_clientIP() should return public IP.
Instead we see the private NAT address.

What version of the product are you using? On what operating system?
All versions with JavaScript API support. Platform independent.

Please provide any additional information below.
Fix is known and will be committed as a patch shortly.

Original issue 50 created by aaronmbr on 2010-06-18T23:21:14.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in client.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. Run above steps after executing setup_ethtool script and compare results.

What is the expected output? What do you see instead?
Inbound throughput should be around 1Gbit but throughput reported is around 700MBit if we run the setup_ethtool script at the NDT server. When this script is not executed throughput reported is around 940Mbit.

What version of the product are you using? On what operating system?
NDT server 3.6.3
Client : Linux (Ubuntu 9)

Please use labels and text to provide additional information.

Original issue 14 created by aaronmbr on 2010-02-27T01:36:41.000Z:

What steps will reproduce the problem?
1.Open the browser
2.Click “Start” button in the applet.
3.Click “Option” button

What is the expected output? What do you see instead?
Expected Output:
Test should be started and all applet buttons should be in disable mode
once test is started.

Observed Output:

After opening the Java client page in Browser, when we run the test first
time(Click Start Button),during the test all buttons in applet are
displaying properly(in disable mode).

But 2nd time onwords(Click Start Button), during execution of the test all
buttons captions in Applet are not displaying. After completion of test,
button captions are displaying.

Browser : Forefox 3.0(Linux)

NDT Version : 3.6.0

Original issue 49 created by aaronmbr on 2010-06-18T23:19:38.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in client.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. Check the RTT reported in the statistics window.

What is the expected output? What do you see instead?
RTT displayed in the statistics window should be matched with the RTT showed by the ping command but RTT displayed by the NDT is more.

What version of the product are you using? On what operating system?
NDT server 3.6.3
Client : Linux (Ubuntu 9)

Please use labels and text to provide additional information.

Original issue 12 created by aaronmbr on 2010-02-27T01:34:50.000Z:

What steps will reproduce the problem?
1.Open the browser
2.Click “Start” button in the applet.
3.Observe test is started or not.

What is the expected output? What do you see instead?
Expected Output:
Applet should conenct to server and should complete the tests and display
the results in applet window.

Observed Output:
1.Size of the Statistics Pop up window is displaying differently for 1st
time test(Click Start Button) and 2nd time test onwords(Click Start Button).

2.Size of the “Close” and “Copy” buttons in Statistics Pop up window is
displaying differently for 1st time test(Click Start Button) and 2nd time
test onwords(Click Start Button).

Browser : Firefox(Linux)

3.“Copy” button in Statistics Pop up window caption is not displaying at
the 1st time test(Click Start Button).

4."Copy" button in Statistics Pop up window is displaying in disable mode
always.

5.Statistics Pop up window is allowing to edit and deleting the displayed
results.

Browser : Firefox 3.0(Linux&MAC),Safari(Mac),Chrome(Linux,Windows) and
IE(Windows & Linux).

6.Statistics,More Details & report Problem buttons are displaying in
disable mode but they are in enable mode only. When clicking those buttons
their respective pop up windows are displaying.
Browser: Firefox .30 in MAC OS

NDT Version : 3.6.0

Original issue 25 created by aaronmbr on 2010-03-04T00:06:54.000Z:

What steps will reproduce the problem?
1.Start test with 100 clients simultaneously connect to server
2.check the messages displayed by 100 clients

What is the expected output? What do you see instead?

Expected Output:
when multiple clients are connecting to the server then wait flag sequence
should be as below.

For 1-20 Clients receiving wait flag as 0
For 21-40 Clients receiving wait flag as 1
For 41-60 Clients receiving wait flag as 2
For 61-80 Clients receiving wait flag as 3
For 81-100 Clients receiving wait flag as 4
For 101- XXX Clients receiving wait flag as 9988

Observed Output:

we have observed the below sequence more frequently.

For 1-20 Clients receiving wait flag as 0
For 21-40 Clients receiving wait flag as 1
For 41-60 Clients receiving wait flag as 2
For 61-80 Clients receiving wait flag as 3
For 81- 100 Clients receiving wait flag as 9988
we have observed that for 100 clients 79 clients are completing the tests
successfully and 21 clients are getting the message as “Server Busy: Too
many clients waiting in server queue. Please try again later.”.

But we have observed the below sequence few times.
For 1-20 Clients receiving wait flag as 0
For 21-40 Clients receiving wait flag as 1
For 41-60 Clients receiving wait flag as 2
For 61-80 Clients receiving wait flag as 3
For 81-82 Clients receiving wait flag as 4
For 82-100 Clients receiving wait flag as 9988

Original issue 5 created by aaronmbr on 2010-02-27T01:24:03.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the following tcp paramters in both server and client.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "0" > /proc/sys/net/ipv4/tcp_ecn
4. Click the start button to start the test and wait till client complete
   the test.
5. In the more details window server and client sizes are displayed 1283K
   and 1852K. But the throughput limit is displayed as 3801Mbps with server
   and 2742Mbs with client.

What is the expected output? What do you see instead?
Since it is showing client is having more buffer client throughput limit
should be more that that server throughput limit displaced.

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

Original issue 40 created by aaronmbr on 2010-03-10T20:41:20.000Z:

Blanket report covering issue 1, issue 10, issue 11, issue 20, issue 22, issue 23, issue 24, and issue
25.. Synopsis:

The server should be able to handle multiple clients at a time. The clients will enter in queues and
be served based on arrival time. The testing has revealed that:

• not all clients enter into the correct queues
• wait time varies
• Zombie process management is needed
• Server may need to be more robust.

See individual bug reports for details

Original issue 27 created by aaronmbr on 2010-03-05T18:16:13.000Z:

Window Scaling option reported ON even it is disabled in both server and
client.

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Set the TCP Parameters of client and server as bellow.
   echo 1 > /proc/sys/net/ipv4/tcp_moderate_rcvbuf
   echo 4194304 > /proc/sys/net/core/wmem_max
   echo 4194304 > /proc/sys/net/core/rmem_max
   echo "4096 87380 4194304" > /proc/sys/net/ipv4/tcp_rmem
   echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
   echo "1" > /proc/sys/net/ipv4/tcp_timestamps
   echo "1" > /proc/sys/net/ipv4/tcp_window_scaling
   echo "1" > /proc/sys/net/ipv4/tcp_sack
   echo "1" > /proc/sys/net/ipv4/tcp_ecn
4. Open the browser and open the NDT home page.
5. Start test by clicking Start button and check the results in statistics
   window.

What is the expected output? What do you see instead?
Since the Window scaling is disabled in both server and client and is not
negotiated this option should be reported as OFF but it is reported as ON.

What version of the product are you using? On what operating system?
NDT server 3.6.1
Client : Linux (Fedora 8)

Please provide any additional information below.

Original issue 10 created by aaronmbr on 2010-02-27T01:28:45.000Z:

What steps will reproduce the problem?
1.Start test with 100 clients simulataneously connect to server
2.check the messages displayed by 100 clients

What is the expected output? What do you see instead?
Expected Output:
All 100 clients should be complete the Middlebox,Firewall , Outbound and
InBound tests.

Observed Output:
Some of the clients didn't completed the test and displaying the “Invalid
test sui” message.

NDT Version : 3.6.0

Original issue 2 created by aaronmbr on 2010-02-27T01:17:57.000Z:

What steps will reproduce the problem?
This problem occured when testing fakewww performance (response times)
using J-Meter tool.

1. Create Test plan with Thread group, Http request defaults, Http request
   to load Html page, and 2 HTTP request to load
   applet (since this in browser applet code requested twice).
2. Configure thread group with 1000 clients, ramp-up period 0 and loop
   count 100.
3. Configure Http request defaults with ip address of server and port
   number (7123)
4. Configure Http requests to load NDT server page and other 2 request to
   load the applet code.
5. Add View results tree listener for all the three HTTP requests.
6. Run the test
7. Server system hang after some time (this is observerd every time with
   local NDT server and rarely with development server).

What is the expected output? What do you see instead?
All clients should complete the transactions with the server. But server is
hanging.

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)
J-Meter : Ver 2.3.4

Please provide any additional information below.

Original issue 23 created by aaronmbr on 2010-03-02T22:29:16.000Z:

What steps will reproduce the problem?
1.Start test with 100 clients simulataneously connect to server
2.check the messages displayed by 100 clients

What is the expected output? What do you see instead?
Expected Output:

All 100 clients should be complete the Middlebox,Firewall , Outbound and
InBound tests.

Observed Output:
Some clients are exiting from the stage at waiting for wait flag data from
the server. After client is get connected with the server server should
send the wait flag 0,1,2,3,4 or 9988 which should be read by the client.
But client is exiting with out reading the wait flag data from server and
displaying "Server busy: Please wait 30 seconds for previous test to
finish". This occurring when tested with multiple clients connect to server
at the same time.
Normally clients should exit after completing the test or if they receive
wait flag other than 0,12,3 or 3 only. But in this case clients did not
received any wait flag before they exit.

NDT Version : 3.6.1
OS : Linux

Original issue 3 created by aaronmbr on 2010-02-27T01:20:41.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Open the browser and connect to the server.
4. Click the start button to start the test and wait till client complete
   the test.
5. Slowest link displayed as (slowest link as "622 Mbps OC-12) insted of
   1GBit also the inbound and outbound throughput is showed more than 900MB.

What is the expected output? What do you see instead?
Slowes Link should be 1GBit ethernet and throughput should be less then the
slowest link.

What version of the product are you using? On what operating system?
NDT server 3.6.0
Client : Linux (Ubuntu 9)

Please provide any additional information below.

In more details window also it is showing link type as OC-12 and GigE.

Original issue 16 created by aaronmbr on 2010-02-27T01:38:03.000Z:

What steps will reproduce the problem?
1.Open the browser
2.Check the links provided in the html page.

What is the expected output? What do you see instead?
Expected Output:

1.Link is not working for "National Science Foundation – VA(USA)"
Browser : Firefox 3.0, Chrome(Linux,MAC & Windows),IE(windows)

2." tester" link under the category of "Other bandwidth testing sites:" is
not working.
Browser :

1. "bt2kfast" link under the category of "Other bandwidth testing sites:"
   is not working.
   Browser :
2. "bandwidth tester" link under the category of "Other bandwidth testing
   sites:" is not working.
   Browser :
   6.ORNL and UT link is not working.

Browser : Firefox 3.0, Chrome(Linux,MAC & Windows),IE(windows)

Original issue 33 created by aaronmbr on 2010-03-05T23:48:04.000Z:

What steps will reproduce the problem?

1. Connect server and client in a 1GBit Switch
2. Start both servers (fakewww and web100)
3. Click the start button to start the test and wait till client complete
   the test.
4. Click the report button on the applet.

What is the expected output? What do you see instead?

When the report button is clicked it should invoke the mail but it is not
opening the mail to send the report. When we click the report problem link
in the page
mail get opened to send the report.

What version of the product are you using? On what operating system?
NDT server 3.6.1
Client : MAC (10.5)
Browser: Firefox (3.0)

Please provide any additional information below.

Original issue 19 created by aaronmbr on 2010-02-27T01:54:27.000Z:

What steps will reproduce the problem?
1.pass the adimistratorview parameter to server as runtime parameter.

What is the expected output? What do you see instead?
Expected Output:
Server should accept the runtime parameter “administrator view” and
initilize the server with received runtime parameter.

Observed Output:

web100 Server is crashing when Administrator view parameter is passing as
an argument.

web100 server is crashing when server is invoking with administrator view
argument. This parameter is passing through config file as below mentioned.

web100srv -4 -c ServConf.txt.

Log file is attached.

Original issue 37 created by aaronmbr on 2010-03-09T13:22:13.000Z:

For UNINETT's deployment of NDT, we've translated the applet into Norwegian. We've finally
gotten around to separating the texts out into a Java resource files. Patch based on ndt-3.5.6 is
attached. The resource files are included in the jar archive, so deployment is easy.

The default is English. If you want to display a different language, add this to the
applet tag:

<PARAM NAME="language" VALUE="nb">
<PARAM NAME="country" VALUE="NO">

If you want to translate to, e.g. French in France, make a file

Tcpbw100_msgs_fr_FR.properties, and go ahead with translation. In this case, the tag would be:

<PARAM NAME="language" VALUE="fr">
<PARAM NAME="country" VALUE="FR">

Accented characters must be entered as unicode literals.

Original issue 26 created by aaronmbr on 2010-03-05T16:54:11.000Z:

What steps will reproduce the problem? (SITUATIONAL)

1. http://broadband.gov/qualitytest/mlabs.htm
2. Test progress continued to well above 100% and took more than a minute
3. Results showed Protocol error! Expected test results, got: ���instead

What is the expected output? What do you see instead? The characters ���
would be replaced with something useful.

What version of the product are you using? On what operating system? All
M-Lab platforms that the FCC can access are at 3.6.1. This tester was
using WinXP and Firefox.

Please provide any additional information below.
Obviously, this will reproduce only in those situations where Protocol
Error occurs.

See attached image.

Original issue 41 created by aaronmbr on 2010-03-10T20:58:28.000Z:

Blanket report covering issue 3 ,issue 4 ,issue 5 ,issue 6 ,issue 7, issue 8, issue 9, issue 17, issue
27, issue 29, issue 30, issue 31, and issue 34. Synopsis:

Local testing on a very controlled environment (network, servers, host settings) shows that the
heuristics in many of the algorithms are not accurate. In general:

• MSS calculation
• Buffer sizes
• Detection of the environment (link speed)

See individual bug reports for details.