[or-cvs] [torperf/master] Add a script to print the chosen path

[sebastian at torproject.org]

commit f59100c6ee87cec248fd9d66c3201898c3df5846
Author: Sebastian Hahn <sebastian at torproject.org>
Date:   Thu Sep 9 21:43:32 2010 +0200

    Add a script to print the chosen path
    
    This script prints out the relays used to fetch a resource, so that the
    resulting speed can be correlated with the relays that were picked while
    measuering.
---
 extra_stats.py     |   63 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 measurements-HOWTO |   41 +++++++++++++++++++++++++++++++++-
 2 files changed, 103 insertions(+), 1 deletions(-)

diff --git a/extra_stats.py b/extra_stats.py
new file mode 100644
index 0000000..3da4c4e
--- /dev/null
+++ b/extra_stats.py
@@ -0,0 +1,63 @@
+import sys, time, math
+import TorCtl.TorCtl as TorCtl
+
+HOST = "127.0.0.1"
+
+class WriteStats(TorCtl.PostEventListener):
+  def __init__(self, port, filename):
+    TorCtl.PostEventListener.__init__(self)
+    self._port = int(port)
+    self._filename = filename
+    self._conn = None
+
+
+  def connect(self):
+    self._conn = TorCtl.connect(HOST, self._port)
+
+  def setup_listener(self):
+    self._conn.set_events(["Stream"])
+    self._conn.add_event_listener(self)
+
+  def stream_status_event(self, event):
+    if event.source == 'EXIT' or \
+       event.status == 'DETACHED':
+      for circs in self._conn.get_info( "circuit-status").itervalues():
+        for circ in circs.split("\n"):
+          entry = circ.split(" ")
+          if entry[0] == str(event.circ_id):
+            if event.source == 'EXIT':
+              self.write_result(True, event.arrived_at, entry[2].split(","))
+            else:
+              self.write_result(False, event.arrived_at, entry[2].split(","))
+
+  def write_result(self, successful, now, relays):
+    success = ""
+    if successful:
+      success = "ok"
+    else:
+      success = "error"
+
+    statsfile = open(self._filename, 'a')
+    statsfile.write("%s %d %s\n" % (success, math.floor(now), " ".join(relays)))
+    statsfile.close()
+
+def main():
+  if len(sys.argv) < 3:
+    print "Bad arguments"
+    sys.exit(1)
+
+  port = sys.argv[1]
+  filename = sys.argv[2]
+
+  stats = WriteStats(port, filename)
+  stats.connect()
+  stats.setup_listener()
+  try:
+    while True:
+      time.sleep(500)
+  except:
+    pass
+
+if __name__ == '__main__':
+  main()
+
diff --git a/measurements-HOWTO b/measurements-HOWTO
index 3c6bf8a..bb236b8 100644
--- a/measurements-HOWTO
+++ b/measurements-HOWTO
@@ -49,6 +49,8 @@ MaxCircuitDirtiness 1 minute
 UseEntryGuards 0
 RunAsDaemon 1
 Log notice file log
+ControlPort 10020
+CookieAuthentication 1
 EOF
 
 $ cat <<EOF >> torclient1mb/torrc
@@ -57,6 +59,8 @@ SocksPort 9021
 UseEntryGuards 0
 RunAsDaemon 1
 Log notice file log
+ControlPort 10021
+CookieAuthentication 1
 EOF
 
 $ cat <<EOF >> torclient5mb/torrc
@@ -65,15 +69,29 @@ SocksPort 9022
 UseEntryGuards 0
 RunAsDaemon 1
 Log notice file log
+ControlPort 10022
+CookieAuthentication 1
 EOF
 
-Write a start script to start the Tor clients and execute it:
+ControlPort is configured so that we can attach a small controller
+that is used to write additional statistics about the chosen paths
+
+Now we need to get TorCtl:
+
+$ git clone git://git.torproject.org/pytorctl TorCtl
+
+Write a start script to start the Tor clients and the controllers,
+and execute it:
 
 $ cat <<EOF >> start-tors
 #!/bin/bash
 cd ~/torperf/torclient50kb && tor -f ~/torperf/torclient50kb/torrc
 cd ~/torperf/torclient1mb && tor -f ~/torperf/torclient1mb/torrc
 cd ~/torperf/torclient5mb && tor -f ~/torperf/torclient5mb/torrc
+sleep 5
+cd ~/torperf/torclient50kb && python ../extra_stats.py 10020 ~/torperf/50kb.extradata
+cd ~/torperf/torclient1mb && python ../extra_stats.py 10021 ~/torperf/1mb.extradata
+cd ~/torperf/torclient5mb && python ../extra_stats.py 10022 ~/torperf/5mb.extradata
 EOF
 
 $ chmod a+x start-tors
@@ -126,6 +144,27 @@ startsec startusec
   datacompletesec datacompleteusec
   writebytes readbytes
 
+The 50kb.extradata, 1mb.extradata, and 5mb.extradata files should
+accumulate lines like this (linebreaks for formatting reasons):
+
+ok 1284059486 $ED16C4DD8E9BCACDE829E6B6571B58095383897A=PPrivCom030
+  $80818E35A46DEED6889818ADC596404AB94E392A=Pandora14
+  $4C23C8C30C66C87E1C875A949B4841E4305FFC57~Amunet11
+
+with column headers:
+
+ok|error timestamp firsthop secondhop thirdhop [fourthhop ...]
+
+The first column indicates if this circuit was actually used to fetch
+the data or if Tor chose a different one because this one was
+problematic. For every error entry there should be a following ok
+entry, unless the network of the torperf instance is totally dead or
+the resource it wants to fetch is unavailable.
+
+The timestamp will NOT necessarily match one from the .data file
+exactly, because additional processing is necessary. You should allow
+for up to a minute of slack when matching the two files up.
+
 If everything works, you might want to let your system start these Tor
 clients on system startup. On Debian, this can be done using a crontab
 entry, too: