Iperf hangs


I have been having problems with 'too many orphaned sockets', and after watching the count for several days I never seen
A lot (mostly 0 and < 5).

However while watching the machine I also notice that at times there are several (usually 8) connections from iepmbw.bnl.org to port 5000 on iepm-bw.slac.stanford.edu.

I just ran a multistream iperf from the command line on the bnl node to the slac node and it appears to hang when done and not return to the command line. I waited over a minute.  Here is the output:

iepm --at-- iepmbw ~/v3src]$ /home/iepm/v3src/bin/iperf -c 134.79.240.38 -p 5000 -i 5 -t 20   -w 8M  -P 8
------------------------------------------------------------
Client connecting to 134.79.240.38, TCP port 5000
TCP window size: 16.0 MByte (WARNING: requested 8.00 MByte)
------------------------------------------------------------
[ 10] local 192.203.218.61 port 33094 connected with 134.79.240.38 port 5000
[  4] local 192.203.218.61 port 33088 connected with 134.79.240.38 port 5000
[  3] local 192.203.218.61 port 33087 connected with 134.79.240.38 port 5000
[  5] local 192.203.218.61 port 33089 connected with 134.79.240.38 port 5000
[  6] local 192.203.218.61 port 33090 connected with 134.79.240.38 port 5000
[  7] local 192.203.218.61 port 33091 connected with 134.79.240.38 port 5000
[  8] local 192.203.218.61 port 33092 connected with 134.79.240.38 port 5000
[  9] local 192.203.218.61 port 33093 connected with 134.79.240.38 port 5000
[  5]  0.0- 5.0 sec  34.0 MBytes  57.0 Mbits/sec
[  7]  0.0- 5.0 sec  19.8 MBytes  33.2 Mbits/sec
[  9]  0.0- 5.0 sec  32.3 MBytes  54.1 Mbits/sec
[  8]  0.0- 5.0 sec  23.8 MBytes  39.9 Mbits/sec
[  4]  0.0- 5.0 sec  38.4 MBytes  64.4 Mbits/sec
[  8]  5.0-10.0 sec  16.4 MBytes  27.5 Mbits/sec
[  7]  5.0-10.0 sec  76.8 MBytes    129 Mbits/sec
[  5]  5.0-10.0 sec  47.9 MBytes  80.3 Mbits/sec
[  4]  5.0-10.0 sec  53.9 MBytes  90.4 Mbits/sec
[  9]  5.0-10.0 sec  20.2 MBytes  33.9 Mbits/sec
[  7] 10.0-15.0 sec  68.2 MBytes    114 Mbits/sec
[  4] 10.0-15.0 sec  43.6 MBytes  73.1 Mbits/sec
[  5] 10.0-15.0 sec  42.2 MBytes  70.9 Mbits/sec
[  9] 10.0-15.0 sec  20.6 MBytes  34.6 Mbits/sec
[  8] 10.0-15.0 sec  36.9 MBytes  62.0 Mbits/sec
[  4] 15.0-20.0 sec  46.1 MBytes  77.3 Mbits/sec
[  4]  0.0-20.3 sec    182 MBytes  75.3 Mbits/sec
[  8] 15.0-20.0 sec  37.5 MBytes  62.9 Mbits/sec
[  8]  0.0-20.3 sec    115 MBytes  47.4 Mbits/sec
[  5] 15.0-20.0 sec  38.3 MBytes  64.2 Mbits/sec
[  5]  0.0-20.4 sec    162 MBytes  66.8 Mbits/sec
[SUM]  0.0-20.4 sec    607 MBytes    250 Mbits/sec
[  9] 15.0-20.0 sec  28.3 MBytes  47.5 Mbits/sec
[  9]  0.0-20.7 sec    101 MBytes  41.0 Mbits/sec
[  7] 15.0-20.0 sec  13.6 MBytes  22.8 Mbits/sec
[  7]  0.0-21.2 sec    178 MBytes  70.4 Mbits/sec
[iepm --at-- iepmbw ~/v3src]$  I did a <ctrl c> to get out of it.

Does anyone have any ideas what may be causing this?

I am running the iperf again, and it is hung before it finishes.  I get the following output:
[iepm --at-- iepmbw bin]$ /home/iepm/v3src/bin/iperf -c 134.79.240.38 -p 5000 -i 5 -t 20 -w 8M -P 8
------------------------------------------------------------
Client connecting to 134.79.240.38, TCP port 5000
TCP window size: 16.0 MByte (WARNING: requested 8.00 MByte)
------------------------------------------------------------
[ 10] local 192.203.218.61 port 34526 connected with 134.79.240.38 port 5000
[  3] local 192.203.218.61 port 34519 connected with 134.79.240.38 port 5000
[  4] local 192.203.218.61 port 34520 connected with 134.79.240.38 port 5000
[  5] local 192.203.218.61 port 34521 connected with 134.79.240.38 port 5000
[  8] local 192.203.218.61 port 34524 connected with 134.79.240.38 port 5000
[  7] local 192.203.218.61 port 34523 connected with 134.79.240.38 port 5000
[  9] local 192.203.218.61 port 34525 connected with 134.79.240.38 port 5000
[  6] local 192.203.218.61 port 34522 connected with 134.79.240.38 port 5000
[  8]  0.0- 5.0 sec  24.3 MBytes  40.7 Mbits/sec
[  7]  0.0- 5.0 sec  24.5 MBytes  41.1 Mbits/sec
[  7]  5.0-10.0 sec  20.4 MBytes  34.2 Mbits/sec
[  8]  5.0-10.0 sec  15.8 MBytes  26.5 Mbits/sec
[  7] 10.0-15.0 sec  33.0 MBytes  55.3 Mbits/sec
[  8] 10.0-15.0 sec  24.4 MBytes  40.9 Mbits/sec
[  8] 15.0-20.0 sec  32.5 MBytes  54.5 Mbits/sec
[  8]  0.0-20.3 sec  96.9 MBytes  40.1 Mbits/sec
[  7] 15.0-20.0 sec  37.4 MBytes  62.8 Mbits/sec
[  7]  0.0-20.4 sec    115 MBytes  47.5 Mbits/sec

And it just hangs there while I see on the SLAC end
48cal --at-- iepm-bw: netstat -a | grep bnl
tcp        0      0 iepm-bw.slac.stanford:33962 iepmbw.bnl.org:ssh          ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34519        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34526        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34525        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34522        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34520        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:34521        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:33087        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:33094        ESTABLISHED
tcp        0      0 iepm-bw.slac.stanford.:5000 iepmbw.bnl.org:33090        ESTABLISHED

I eventually killed the iperf command runnning from BNL, and the sockets are still allocated
Several minutes later on the slac node.


Any ideas?

Thanks Connie
Connie Logg, Network Analyst
Stanford Linear Accelerator Center
ph: 650-926-2879 
"Happiness is found along the way, not at the end of the road, and 'IF' is the middle word in life."



Other Mailing lists | Author Index | Date Index | Subject Index | Thread Index