[tahoe-lafs-trac-stream] [tahoe-lafs] #1456: High latency for 'tahoe get' if 'tahoe put' in parallel

[tahoe-lafs]

#1456: High latency for 'tahoe get' if 'tahoe put' in parallel
-------------------------+-------------------------------------------------
     Reporter:  T_X      |      Owner:  T_X
         Type:  defect   |     Status:  new
     Priority:           |  Milestone:  undecided
  critical               |    Version:  1.8.2
    Component:  code     |   Keywords:  download upload latency performance
   Resolution:           |  gateway vm kvm vpn trickle
Launchpad Bug:           |
-------------------------+-------------------------------------------------

Comment (by T_X):

 I'm already using statements like
 {{{
 print "lineReceived %s" % (time.asctime( time.localtime(time.time()) ),)
 }}}
 and this printed timestamp is usually already too late, that's why I guess
 the issue has to be before the lineReceived call.

 Hmm, I'm usually running 'tahoe run' instead of 'tahoe start' for the 2nd
 node, the gateway node doing the put and get, because trickle does not
 seem to work for 'tahoe start'. But for 'tahoe run' this profiler does not
 seem to be available.

 Therefore I started to do my own, manual profiling... Basically I had a
 look at a call trace of lineReceived and tried surrounding various
 function calls in there with print's and timestamps. However, I couldn't
 find a function consuming this awful lot of time yet. twisted's tcp.py's
 Connection's doRead never needs more than a second.

-- 
Ticket URL: <http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1456#comment:12>
tahoe-lafs <http://tahoe-lafs.org>
secure decentralized storage