Re: clock jump error persists into 0.2.0.25-rc

[Scott Bennett <bennett@xxxxxxxxxx>]

     On Fri, 25 Apr 2008 00:53:36 -0400 phobos@xxxxxxxxxx wrote:
>On Thu, Apr 24, 2008 at 02:43:08PM -0500, bennett@xxxxxxxxxx wrote 4.6K bytes in 52 lines about:
>:      I compiled and installed 0.2.0.25-rc to try it out, and I'm sad to write,
>: it still has the "clock jump" problem and isn't usable:
>
>I'm starting to think this is some interaction between your hardware and
>Tor.  I setup a freebsd 6.3-stable server on my private tor network.  It

     That's possible.  Nevertheless, with 0.2.0.19-alpha and earlier, the
speculative interaction didn't happen, so it would seem to involve changes
made to tor since then.  That should limit the number of potential trouble
spot(s) to examine.  Even better, although I don't think I tried
0.2.0.20-rc, I know that 0.2.0.21-rc did have the problem, which further
reduces the changes made, though there may not have been any additional
changes to time-related code in tor after that, so later release changes
 might not make a difference in that regard.

>hasn't experienced this clock jumping problem.  The hardware is a quad
>xeon with HT enabled, specifically, "Intel(R) Xeon(TM) CPU 2.80GHz".  I
>built my own tor from source. 

     I'll ask out of curiousity, do you have NumCPUs 8 in your torrc?  I have
NumCPUs 2 in mine, a 3.4 GHz P4 with HT enabled.  I also build tor from source.
(With MAKEFLAGS set to -j5, it only takes a minute or so to build, IIRC.:-)
>
>I may put it online in the real tor network and see if things change.
>As of right now, I can't replicate the problem, so I can't suggest a
>fix.  Perhaps run with debug logs for a bit or truss tor to see what's
>going on when Tor says the clock jumps.

     Debug-level logs tend to be humongous.  Haven't used truss before, but it
looks like its output could be about that voluminous, too.  Is there somewhere
you'd like the results sent if I do run it that way long enough to collect some
evidence?  In the past when tor developers have asked me to do that for some
problem, I was unable to get a response to that question after I'd collected
the log output for them, so I'm hesitant to tie up that much space again unless
I know in advance what to do with it.  


                                  Scott Bennett, Comm. ASMELG, CFIAG
**********************************************************************
* Internet:       bennett at cs.niu.edu                              *
*--------------------------------------------------------------------*
* "A well regulated and disciplined militia, is at all times a good  *
* objection to the introduction of that bane of all free governments *
* -- a standing army."                                               *
*    -- Gov. John Hancock, New York Journal, 28 January 1790         *
**********************************************************************