On 30 December 2011 07:50, pc <paul.cro...@softwareag.com> wrote: > Hallo Miguel, > > Yes -- if TRAIM is configured in a particular way, the PPS > "pulse [is] active only when Time RAIM algorithm confirms > time solution error is within the user defined alarm limit". > > Can you poll the current TRAIM setup and status > with an @@En command and post the output? >
Hi Paul! I suspected TRAIM could have something to do with this and so I disabled it in the configuration file # cat /etc/ntp.oncore.0 MODE 1 LAT 40 55 13.1082 LONG -8 29 36.7247 HT 319.03 M TRAIM NO DELAY 25 NS CLEAR MASK 30 HARDPPS I even set mask to 30 degrees. Once I disabled TRAIM I am getting a good behavior (no more jumps in the clocks file and in the offset in the loop). Here are some stats for the past 3 days... This is the loop offset (in absolute value): # cat /var/log/ntp/loop.20111227 | /mnt/script.sh NTP Loopfile Analysis --------------------- 0 us: 1940 samples (35.92 %) (35.92 %) 1 us: 1460 samples (27.04 %) (62.96 %) 2 us: 604 samples (11.19 %) (74.15 %) 3 us: 263 samples (4.87 %) (79.02 %) 4 us: 286 samples (5.30 %) (84.31 %) 5 us: 351 samples (6.50 %) (90.81 %) 6 us: 216 samples (4.00 %) (94.81 %) 7 us: 181 samples (3.35 %) (98.16 %) 8 us: 88 samples (1.63 %) (99.79 %) 9 us: 11 samples (0.20 %) (100.00 %) 5400 / 5400 samples evaluated # cat /var/log/ntp/loop.20111228 | /mnt/script.sh NTP Loopfile Analysis --------------------- 0 us: 2395 samples (44.35 %) (44.35 %) 1 us: 1195 samples (22.13 %) (66.48 %) 2 us: 330 samples (6.11 %) (72.59 %) 3 us: 195 samples (3.61 %) (76.20 %) 4 us: 168 samples (3.11 %) (79.31 %) 5 us: 163 samples (3.02 %) (82.33 %) 6 us: 267 samples (4.94 %) (87.28 %) 7 us: 318 samples (5.89 %) (93.16 %) 8 us: 172 samples (3.19 %) (96.35 %) 9 us: 80 samples (1.48 %) (97.83 %) 10 us: 63 samples (1.17 %) (99.00 %) 11 us: 44 samples (0.81 %) (99.81 %) 12 us: 10 samples (0.18 %) (99.99 %) 5400 / 5400 samples evaluated # cat /var/log/ntp/loop.20111229 | /mnt/script.sh NTP Loopfile Analysis --------------------- 0 us: 1931 samples (35.76 %) (35.76 %) 1 us: 1357 samples (25.13 %) (60.89 %) 2 us: 621 samples (11.50 %) (72.39 %) 3 us: 344 samples (6.37 %) (78.76 %) 4 us: 355 samples (6.57 %) (85.33 %) 5 us: 292 samples (5.41 %) (90.74 %) 6 us: 202 samples (3.74 %) (94.48 %) 7 us: 192 samples (3.56 %) (98.03 %) 8 us: 75 samples (1.39 %) (99.42 %) 9 us: 12 samples (0.22 %) (99.64 %) 10 us: 15 samples (0.28 %) (99.92 %) 11 us: 4 samples (0.07 %) (100.00 %) 5400 / 5400 samples evaluated And for today (at least until 12:30Z): # cat /var/log/ntp/loop.20111230 | /mnt/script.sh NTP Loopfile Analysis --------------------- 0 us: 1045 samples (37.12 %) (37.12 %) 1 us: 845 samples (30.02 %) (67.14 %) 2 us: 106 samples (3.77 %) (70.90 %) 3 us: 32 samples (1.14 %) (72.04 %) 4 us: 100 samples (3.55 %) (75.59 %) 5 us: 247 samples (8.77 %) (84.37 %) 6 us: 170 samples (6.04 %) (90.41 %) 7 us: 150 samples (5.33 %) (95.73 %) 8 us: 93 samples (3.30 %) (99.04 %) 9 us: 17 samples (0.60 %) (99.64 %) 10 us: 6 samples (0.21 %) (99.85 %) 11 us: 2 samples (0.07 %) (99.92 %) 12 us: 2 samples (0.07 %) (99.99 %) 2815 / 2815 samples evaluated The Oncore is outputting every second: # wc -l /var/log/ntp/clock.2011122[7,8,9] 86401 /var/log/ntp/clock.20111227 86401 /var/log/ntp/clock.20111228 86401 /var/log/ntp/clock.20111229 The extra line over 86400 is the one that contains the leap second information. Unfortunately I don't have the software to query the Oncore unit as you suggest. I only have a FreeBSD machine and am using the driver from the NTP distribution. Any suggestions? Many thanks Paul! Cheers, Miguel _______________________________________________ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
