Once it has reached a SLAVE state, sfptpd generates output describing the current state of the synchronization.
By default sfptpd will output the offset data to
stdout. You can redirect stats_log output to
file using the configuration file stats_log
parameter.
When logging the stats_log to file, it is possible to override this using the -v option on the sfptpd command line to cause stats to display on stdout.
The following examples are from a PTP sync module on a slave server. The timestamp at the start of the line has been omitted. Other sync modules produce similar output.
[ptp1:gm->phc0(enp1s0f0)], offset: 0.500, freq-adj: -663.353, in-sync: 1,
one-way-delay: 7534.500, grandmaster-id: 00a0:69ff:fe0c:2eb5
[phc0(enp1s0f0/enp1s0f1)->system], offset: 2.812, freq-adj: 46378.715,
in-sync: 1| Parameter | Description |
|---|---|
[ptp1:gm-> phc0(enp1s0f0)]
|
A line of output generated for measurements between the external master clock and the Local Reference Clock:
|
offset
|
The current offset (nanoseconds) between the master clock and the slave clock identified at the start of the line. Immediately following startup this is expected to be a large value, but will gradually decrease until it settles to its lowest value. Synchronization can typically take between 15-30 minutes. |
freq-adj
|
The current rate (PPB) at which the clock is being disciplined by sfptpd. This value is stored in the freq-correction file for this clock every 60 seconds. |
in-sync
|
The in-sync flag will be 1 when the offset between master and slave clocks is
below the value of the The local_sync_threshold when using software timestamping is 100µs. The in-sync flag will be 0 before the above condition is true. The in-sync flag will be 0 if the offset becomes greater than the local_sync_threshold. The in-sync flag will change to 0 if an alarm condition exists on the server to indicate problems in the PTP network, such as PTP messages not being sent or received by the slave server. Check the topology file for current alarms status. |
one-way-delay
|
The current one-way-delay (nanoseconds) between master and slave servers. This value should not be zero, but, once the server is synchronized, it should remain fairly stable:
Check the topology file for current alarms status. |
| parent-id | UUID of the upstream master clock. This can be a boundary clock or the GM Master clock. |
gm-id
|
Master clock UUID derived from its MAC address. |
[phc0(enp1s0f0/enp1s0f1)->system]
|
A line of output generated for measurements between the Local Reference Clock and the server system clock:
|