Changes between Initial Version and Version 1 of PlasticSlices/Continuation/Round4


Ignore:
Timestamp:
03/07/12 10:49:16 (10 years ago)
Author:
Josh Smift
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • PlasticSlices/Continuation/Round4

    v1 v1  
     1[[PageOutline]]
     2
     3= Continuation Round 4 =
     4
     5In this round, we ran plastic-105 and plastic-106 for about three hours, with only GT and Wisconsin, to try to track down some issues with performance in Round 3 that seemed related to GT. In particular, we wanted to find out whether there was a specific problem with plnode1.
     6
     7The raw logs are at http://www.gpolab.bbn.com/plastic-slices/continuation/round-4/.
     8
     9= plastic-105 =
     10
     11[ggw:PlasticSlices/Experiments#SteadyPerfTCP SteadyPerf TCP], using port=5105, time=518400, and this table of client/server pairs:
     12
     13|| '''client'''                   || '''server'''                   || '''server address''' ||
     14|| plnode1.cip.gatech.edu         || wings-openflow-2.wail.wisc.edu || server=10.42.105.95  ||
     15|| plnode2.cip.gatech.edu         || wings-openflow-3.wail.wisc.edu || server=10.42.105.96  ||
     16
     17== One-time prep commands run on each client and server ==
     18
     19{{{
     20sudo yum -y install iperf
     21}}}
     22
     23== Commands run on each server ==
     24
     25{{{
     26server=<ipaddr>
     27nice -n 19 iperf -B $server -p 5105 -s -i 1
     28}}}
     29
     30== Commands run on each client ==
     31
     32{{{
     33server=<ipaddr>
     34nice -n 19 iperf -c $server -p 5105 -t 518400
     35}}}
     36
     37== Results ==
     38
     39Generated with
     40
     41{{{
     42subnet=105
     43for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 5 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done
     44}}}
     45
     46and then edited slightly to remove artifacts (like control characters, my prompt, etc).
     47
     48plnode1.cip.gatech.edu:
     49
     50{{{
     51------------------------------------------------------------
     52Client connecting to 10.42.105.95, TCP port 5105
     53TCP window size: 16.0 KByte (default)
     54------------------------------------------------------------
     55[  3] local 10.42.105.100 port 59778 connected with 10.42.105.95 port 5105
     56[  3]  0.0-13761.8 sec    405 MBytes    247 Kbits/sec
     57}}}
     58
     59plnode2.cip.gatech.edu:
     60
     61{{{
     62------------------------------------------------------------
     63Client connecting to 10.42.105.96, TCP port 5105
     64TCP window size: 16.0 KByte (default)
     65------------------------------------------------------------
     66[  3] local 10.42.105.101 port 58242 connected with 10.42.105.96 port 5105
     67[  3]  0.0-13768.4 sec    482 MBytes    294 Kbits/sec
     68}}}
     69
     70== Analysis ==
     71
     72Less than 300 Kbits/sec is very low.
     73
     74Looking at the server logs shows times when the rate jumped up to one or two Mbits/sec, before dropping back down to double-digt Kbits/sec.
     75
     76The results are similar for the two plnodes, so there probably isn't a problem with just one or the other.
     77
     78= plastic-106 =
     79
     80[ggw:PlasticSlices/Experiments#SteadyPerfUDP SteadyPerf UDP], using port=5106, time=518400, rate=30, and this table of client/server pairs:
     81
     82|| '''client'''                   || '''server'''                   || '''server address''' ||
     83|| plnode1.cip.gatech.edu         || wings-openflow-3.wail.wisc.edu || server=10.42.105.96  ||
     84|| plnode2.cip.gatech.edu         || wings-openflow-2.wail.wisc.edu || server=10.42.105.95  ||
     85
     86== One-time prep commands run on each client and server ==
     87
     88{{{
     89sudo yum -y install iperf
     90}}}
     91
     92== Commands run on each server ==
     93
     94{{{
     95server=<ipaddr>
     96nice -n 19 iperf -u -B $server -p 5106 -s -i 1
     97}}}
     98
     99== Commands run on each client ==
     100
     101{{{
     102server=<ipaddr>
     103nice -n 19 iperf -u -c $server -p 5106 -t 518400 -b 30M
     104}}}
     105
     106== Results ==
     107
     108Generated with
     109
     110{{{
     111subnet=106
     112for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 10 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done
     113}}}
     114
     115and then edited slightly to remove artifacts (like control characters, my prompt, etc).
     116
     117plnode1.cip.gatech.edu:
     118
     119{{{
     120------------------------------------------------------------
     121Client connecting to 10.42.105.96, UDP port 5106
     122Sending 1470 byte datagrams
     123UDP buffer size:   122 KByte (default)
     124------------------------------------------------------------
     125[  3] local 10.42.105.100 port 53151 connected with 10.42.105.96 port 5106
     126[  3]  0.0-13755.2 sec  48.0 GBytes  30.0 Mbits/sec
     127[  3] Sent 35057619 datagrams
     128[  3] Server Report:
     129[  3]  0.0-13754.8 sec  44.2 GBytes  27.6 Mbits/sec  0.078 ms 2767576/35057619 (7.9%)
     130[  3]  0.0-13754.8 sec  252 datagrams received out-of-order
     131}}}
     132
     133plnode2.cip.gatech.edu:
     134
     135{{{
     136------------------------------------------------------------
     137Client connecting to 10.42.105.95, UDP port 5106
     138Sending 1470 byte datagrams
     139UDP buffer size:   122 KByte (default)
     140------------------------------------------------------------
     141[  3] local 10.42.105.101 port 39327 connected with 10.42.105.95 port 5106
     142[  3]  0.0-13756.4 sec  48.0 GBytes  30.0 Mbits/sec
     143[  3] Sent 35068319 datagrams
     144[  3] Server Report:
     145[  3]  0.0-13756.0 sec  44.4 GBytes  27.7 Mbits/sec  0.252 ms 2666045/35068319 (7.6%)
     146[  3]  0.0-13756.0 sec  319 datagrams received out-of-order
     147}}}
     148
     149== Analysis ==
     150
     151Similar packet loss to Round 3, where we saw 7.5% loss.
     152
     153The results are similar for the two plnodes, so there probably isn't a problem with just one or the other.