[Savannah-help-public] Savannah clock is about twenty seconds slow

[Sylvain Beucler]

Hello sysadmins,

Greg suggests that the Savannah clock, which is sync'd with
ns1.gnu.org, is a bit late (~20secs).

I indeed note a difference between by computer and sv:
me# LANG=C date -u +%c
Sun Mar 11 18:28:07 2007
sv# LANG=C date -u +%c
Sun Mar 11 18:27:41 2007

The NTP connection looks good afaics:
savannah:~# ntpdc
ntpdc> peers
     remote           local      st poll reach  delay   offset    disp
=======================================================================
=ns1.gnu.org     199.232.41.3    16 1024    0 0.00000  0.000000 0.00000

So maybe ns1.gnu.org's hour is a bit late.

Can you have a look at it?

Thanks,

-- 
Sylvain


----- Forwarded message from Greg Chicares <address@hidden> -----

To: Greg Chicares <address@hidden>, address@hidden
From: Greg Chicares <address@hidden>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
        rv:1.8.0.10) Gecko/20070216 Firefox/1.5.0.10
Date: Sat, 10 Mar 2007 19:53:34 +0000
Cc: 
Subject: [Savannah-help-public] [sr #105799] Savannah clock is about twenty
        seconds slow


URL:
  <http://savannah.gnu.org/support/?105799>

                 Summary: Savannah clock is about twenty seconds slow
                 Project: Savannah Administration
            Submitted by: chicares
            Submitted on: Saturday 03/10/2007 at 19:53
                Category: Developer CVS
                Priority: 5 - Normal
                Severity: 3 - Normal
                  Status: None
             Assigned to: None
        Originator Email: 
        Operating System: None
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

The cvs server's system clock has always been exactly
correct AFAICT, but in the past several days it has
lagged. Can its former accuracy be restored please?

I'm using ms windows, which has a known issue with
daylight saving time this year, but I think I can prove
that's not the problem. I calibrated my system clock
using a US government time server, then performed the
following three actions as quickly as I could:

(1) First, freshly reload this page:
  http://tycho.usno.navy.mil/cgi-bin/timer.pl
which said:
  Mar. 10, 19:01:35 UTC

(2) Then, in my local development directory:

/c/lmi/src/lmi[0]$date -u +'%Y%m%dT%H:%M:%SZ'; \
 cvs commit -m"Sort forward declarations in ASCII order" xenumtypes.cpp ; \
 date -u +'%Y%m%dT%H:%M:%SZ'; \
 ls -o --full-time xenumtypes.cpp
20070310T19:01:38Z
/sources/lmi/lmi/xenumtypes.cpp,v  <--  xenumtypes.cpp
new revision: 1.20; previous revision: 1.19
Mailing notification to [SUPPRESSED]... sent.
20070310T19:01:40Z
-rw-r--r--    1 chicares    34107 Sat Mar 10 19:01:17 2007 xenumtypes.cpp

The time on the last line above, which comes from the
savannah cvs server AIUI, is eighteen seconds before the
official time in (1), although I did step (1) first.

(3) Then, in a local directory where I mirror cvs:

/c/lmi/checkouts/current[0]$date -u +'%Y%m%dT%H:%M:%SZ'; \
 cvs -z3 update -l lmi; \
 date -u +'%Y%m%dT%H:%M:%SZ'; \
 ls -o --full-time lmi/xenumtypes.cpp
20070310T19:01:39Z
cvs update: Updating lmi
P lmi/xenumtypes.cpp
20070310T19:01:41Z
-rw-r--r--    1 chicares    34107 Sat Mar 10 19:01:40 2007
lmi/xenumtypes.cpp

I believe the time on the last line is correct. It's
five seconds after the "official" time in step (1): a
reasonable (positive) latency.

Now I check cvs through the web interface:
http://cvs.savannah.gnu.org/viewcvs/lmi/lmi/xenumtypes.cpp?rev=1.20&sortby=date&view=log
and it matches (2), which I believe to be incorrect:
  Sat Mar 10 19:01:17 2007 UTC
Comparing that to the "official" time in (1) rules out
any anomalous behavior of my local machine.





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/support/?105799>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/




----- End forwarded message -----