[ntp:questions] Re: using ACTS modem time service

David L. Mills mills at udel.edu
Fri Oct 29 02:16:43 UTC 2004


cipo,

A week or so back I completely ovehauled the refclock_acts.c driver. It 
had picked up many weeds planted by other folks since I wrote the 
original driver 12 years ago. The new driver works very well with both 
ACTS and USNO and supposedly European modem services and automatically 
detects which one. The refclock_usno.c and refclock_ptbacts.c drivers 
are no longer necessary and have been removed from the distribution.

You will need a recent tarball not more than two weeks old. Don't forget 
to set minpoll to 12 and maxpoll to 17. It may take a few calls to start 
increasing the poll interval above 12 (4096 s). When you first start it, 
remove the ntp.drift file and let the clock state machine compute the 
frequency; this speeds up the process and reduces the total call 
traffic. Also, repeat the ACTS number a couple or three times on the 
phones command so it will retry immediately if the call fails.

Dave

cipo wrote:
> We have tried the ACTS modem time service with a standard ntp 
> (ntp-4.2.0,linux-2.4.27,2xP3 and P4 system)
> configuration, but experienced a very strange problem.
> After starting the ntp daemon the modem connects to the service, 
> receives the
> time signals (8-12 pieces), calculates the different ntp values (offset, 
> jitter,
> ...) and hangs up the modem. Immediately after that we receive an error 
> message
> stating that all internal buffers are full ("too many recvbufs allocated 
> (40)") and the ntp daemon consumes
> 100% processor time from that point on. We tried with 2 different 
> modems, and 3
> different modem services (ACTS in the US, PTB in Germany, METAS in 
> Switzerland),
> but the results were the same.
> 
> Has anyone experienced the same problem? We would appreciate any 
> information
> regarding this issue.
> 
> Thans in advance,
> -cipo
> 
> Below is the piece of log from ntpd -dddd:
> ntpd 4.2.0 at 1.1161-r Thu Oct 28 17:42:24 CEST 2004 (5)
> addto_syslog: ntpd 4.2.0 at 1.1161-r Thu Oct 28 17:42:24 CEST 2004 (5)
> addto_syslog: signal_no_reset: signal 13 had flags 4000000
> addto_syslog: set_process_priority: Leave priority alone: priority_done 
> is <2>
> addto_syslog: precision = 1.000 usec
> create_sockets(123)
> addto_syslog: no IPv6 interfaces found
> ...
> init_io: maxactivefd 0
> local_clock: at 0 state 0
> getaddrinfo 127.127.23.0
> getnetnum given 127.127.23.0, got 127.127.23.0 peer_clear: at 0 assoc ID 
> 35812 refid INIT
> refclock_open: fd 5 modem status 0x4126
> refclock_ioctl: fd 5 flags 0x4
> acts: state 0 send 20 ATB1&C1&D2E0L1M1Q0V1
> newpeer: 127.0.0.1->127.127.23.0 mode 3 vers 4 poll 10 18 flags 0x21 0x1 
> ttl 0 key 00000000
> addto_syslog: frequency initialized 111.654 PPM from /etc/ntp.drift
> local_clock: at 0 state 1
> authtrust: keyid 0000ffff life 1
> report_event: system event 'event_restart' (0x01) status 'sync_alarm, 
> sync_unspec, 1 event, event_unspec' (0xc010)
> addto_syslog: input_handler: Processed a gob of fd's in 0.845000 msec
> refclock_gtlin: fd 5 time 3307985312.205591
> refclock_gtlin: fd 5 time 3307985312.205591
> addto_syslog: input_handler: Processed a gob of fd's in 0.008000 msec
> refclock_gtlin: fd 5 time 3307985312.235103 timecode 2 OK
> acts: state 0 timecode 2 OK
> addto_syslog: input_handler: Processed a gob of fd's in 0.008000 msec
> refclock_gtlin: fd 5 time 3307985312.245105
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> auth_agekeys: at 1 keys 1 expired 0
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> refclock_transmit: at 2 127.127.23.0
> acts: state 0 send 7 ATM0DS0
> poll_update: at 2 127.127.23.0 flags 0021 poll 10 burst 0 last 2 next 1026
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> addto_syslog: input_handler: Processed a gob of fd's in 0.013000 msec
> refclock_gtlin: fd 5 time 3307985315.845510
> addto_syslog: input_handler: Processed a gob of fd's in 0.008000 msec
> refclock_gtlin: fd 5 time 3307985315.855509
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> addto_syslog: input_handler: Processed a gob of fd's in 0.010000 msec
> refclock_gtlin: fd 5 time 3307985316.025527 timecode 20 
> T00049531512038#5214
> acts: state 1 timecode 20 T00049531512038#5214
> addto_syslog: input_handler: Processed a gob of fd's in 0.007000 msec
> refclock_gtlin: fd 5 time 3307985316.035529
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> ...
> addto_syslog: input_handler: Processed a gob of fd's in 0.016000 msec
> refclock_gtlin: fd 5 time 3307985349.209241
> addto_syslog: input_handler: Processed a gob of fd's in 0.007000 msec
> refclock_gtlin: fd 5 time 3307985349.219238
> addto_syslog: input_handler: Processed a gob of fd's in 0.009000 msec
> refclock_gtlin: fd 5 time 3307985349.329249 timecode 12 CONNECT 1200
> acts: state 1 timecode 12 CONNECT 1200
> addto_syslog: input_handler: Processed a gob of fd's in 0.007000 msec
> refclock_gtlin: fd 5 time 3307985349.339252
> addto_syslog: input_handler: Processed a gob of fd's in 0.018000 msec
> refclock_gtlin: fd 5 time 3307985349.999328 timecode 78 2004-10-28 
> 22:49:06 MESZ 44430210310320041028204953306-50000504MEZ=CET=UTC+1h*
> acts: state 2 timecode 78 2004-10-28 22:49:06 MESZ 
> 44430210310320041028204953306-50000504MEZ=CET=UTC+1h*
> addto_syslog: input_handler: Processed a gob of fd's in 0.007000 msec
> refclock_gtlin: fd 5 time 3307985350.009325
> ...
> acts: state 3 timecode 78 2004-10-28 22:49:15 MESZ 
> 44430210310320041028204953306-50000506 MESZ=UTC+2h  *
> addto_syslog: input_handler: Processed a gob of fd's in 0.007000 msec
> refclock_gtlin: fd 5 time 3307985355.989992
> addto_syslog: select(): nfound=-1, error: Interrupted system call
> addto_syslog: input_handler: Processed a gob of fd's in 0.023000 msec
> refclock_gtlin: fd 5 time 3307985356.640072 timecode 78 2004-10-28 
> 22:49:16 MESZ 44430210310320041028204953306-50000500PTB-Zeit ueber*
> acts: state 3 timecode 78 2004-10-28 22:49:16 MESZ 
> 44430210310320041028204953306-50000500PTB-Zeit ueber*
> refclock_receive: at 44 127.127.23.0
> peer ACTS_PTB(0) event 'event_reach' (0x84) status 'unreach, conf, 1 
> event, event_reach' (0x8014)
> refclock_sample: n 10 offset -1.479881 disp 0.000000 jitter 1.679621
> clock_filter: n 1 off -1.479881 del 0.000000 dsp 9.348552 jit 0.000001, 
> age 44
> opening filegen (type=2/id=3307910400) 
> "/var/log/ntpstats/peerstats.20041028"
> addto_syslog: input_handler: Processed a gob of fd's in 0.008000 msec
> refclock_gtlin: fd 5 time 3307985356.650067
> addto_syslog: too many recvbufs allocated (40)
> addto_syslog: too many recvbufs allocated (40)
> addto_syslog: too many recvbufs allocated (40)
> 
> 
> 
> 
> 




More information about the questions mailing list