we have been running ptpd2 2.3.1 for a while without issues, but since a few days, after performing some updates (we , the slaves do not sync anymore.
the status file shows
...
Performance : Message RX 121/s, TX 1/s
Announce received : 97
Sync received : 192
Follow-up received : 0
DelayReq sent : 194
DelayResp received : 194
State transitions : 3
PTP Engine resets : 1
so no follow-up messages. however running
tcpdump -i em1 port 320 or port 319
shows sync messages on port 319 and follow-up messages on port 320
the slaves have no iptables
is there a way to check that the sync and followup messages are correct?
(wireshark shows the same sequenceid, the follow is received after the sync, but i noticed that the flags are different: sync has ptp_unicast=false/twostep=true; the followup has flags unicast=true/twostep=false)
did some more debugging, including rebuild using runtime debug and a rebuild from lastest master form github
when running 2.3.1 with -DD, after receiving the sync, the output has a message
(lstn_init) ignord unicast message in multicast mode2
(this is in hybrid)
but the tcpdump output shows that the followup msg was sent mcast, but with the ptp_unicast bit on.
with latest ghmaster code, the followup message does not have the unicast bit set, and now i have follow-up messages again on the 2.3.1 slaves (only the master runs the ghmaster code)
so my question is : this was clearly fixed, is there an intermediate release we can use that has this fix?
on an unrelated note, is the --with-max-unicast-destinations option in configure relevant for hybrid mode?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
hi all,
we have been running ptpd2 2.3.1 for a while without issues, but since a few days, after performing some updates (we , the slaves do not sync anymore.
the status file shows
...
Performance : Message RX 121/s, TX 1/s
Announce received : 97
Sync received : 192
Follow-up received : 0
DelayReq sent : 194
DelayResp received : 194
State transitions : 3
PTP Engine resets : 1
so no follow-up messages. however running
tcpdump -i em1 port 320 or port 319
shows sync messages on port 319 and follow-up messages on port 320
the slaves have no iptables
is there a way to check that the sync and followup messages are correct?
(wireshark shows the same sequenceid, the follow is received after the sync, but i noticed that the flags are different: sync has ptp_unicast=false/twostep=true; the followup has flags unicast=true/twostep=false)
any help is welcome
stijn
current ptpd.conf
clock:drift_file=/var/log/ptpd2_kernelclock.drift
clock:drift_handling=file
clock:set_rtc_on_step=y
clock:step_startup=y
global:cpuaffinity_cpucore=0
global:log_file=/var/log/ptpd2.log
global:log_file_max_files=5
global:log_file_max_size=5000
global:log_status=y
global:statistics_update_interval=5
ptpengine:domain=0
ptpengine:interface=em1
ptpengine:ip_mode=hybrid
ptpengine:log_delayreq_interval=0
ptpengine:panic_mode=y
ptpengine:panic_mode_duration=5
ptpengine:preset=slaveonly
ptpengine:use_libpcap=y
ptpengine:sync_sequence_checking=y
did some more debugging, including rebuild using runtime debug and a rebuild from lastest master form github
when running 2.3.1 with -DD, after receiving the sync, the output has a message
(lstn_init) ignord unicast message in multicast mode2
(this is in hybrid)
but the tcpdump output shows that the followup msg was sent mcast, but with the ptp_unicast bit on.
with latest ghmaster code, the followup message does not have the unicast bit set, and now i have follow-up messages again on the 2.3.1 slaves (only the master runs the ghmaster code)
so my question is : this was clearly fixed, is there an intermediate release we can use that has this fix?
on an unrelated note, is the --with-max-unicast-destinations option in configure relevant for hybrid mode?