I'm a beginner to timesync (do my best to become an expert :) ) and I've just started playing with the daemon. I have some questions concerning ptpd output file. I'm trying to compare it somehow with the ntpd-generated statistics and need some help. I have 2 PC's connected with cross-cable and run daemon:
state: init .....................................................
state: lstn .....................................................
state: mst .....................................................
I guess it won't give me any info when the state is master (?)
My question is what 'owd', 'ofm' and 'var' means (I'm ok with drift ;) ) and how often the 'state...' lines are generated? is it 16sec (-y4)?
Actually, feel free to provide me with as much help as possible as I'm not best at C development (yet ;) )
cheers
Michal
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This is a bug which I have recently fixed in the ptpv2d branch.
Unfortunately it looks like the complete code base has initially simply been copied to create the v2 branch, and after this both code bases have been changed in different ways.
I'll try to discuss with the other developers how to proceed. IMO the best solution would be to change the code base so that utility functions like printing time stamps, or even the servo loop used to discipline the system time, would be available only in a single source file.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just a quick note to say that we have talked via email about this and we are in agreement on creating a single source base going forwards. The initial goal was to have a working v1 and a working v2 before making the single source base. That goal is now nearly achieved. Once both daemons are stable we will create a new daemon, likely to be called v3, which supports both v1 and v2.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi everyone,
I'm a beginner to timesync (do my best to become an expert :) ) and I've just started playing with the daemon. I have some questions concerning ptpd output file. I'm trying to compare it somehow with the ntpd-generated statistics and need some help. I have 2 PC's connected with cross-cable and run daemon:
#1 ./ptpd -c -d -t -p -y4 -s2 -u 192.168.0.2 -f stats.txt (master)
#2 ./ptpd -u 192.168.0.1 -f stats.txt (slave)
master output
state: init .....................................................
state: lstn .....................................................
state: mst .....................................................
I guess it won't give me any info when the state is master (?)
slave output
state: init ...................................................................................................
state: lstn ...................................................................................................
state: slv, owd: 0.0000000000, ofm: 0.0000000000, drift: 0, var: 0 .............
state: slv, owd: 0.0000000000, ofm: 0.0002000000, drift: 200 var: 0 ..........
and so on
My question is what 'owd', 'ofm' and 'var' means (I'm ok with drift ;) ) and how often the 'state...' lines are generated? is it 16sec (-y4)?
Actually, feel free to provide me with as much help as possible as I'm not best at C development (yet ;) )
cheers
Michal
Hello Michal,
owd = one way delay
ofm = offset from master
var = variance
-y4 -> a sync message is sent every 2^4 sec = 16sec
Don't forget to read the provided documentation (HTML page and ptpd_2005_1588_conference_paper.pdf file).
Regards,
Dave.
Dave,
thanks for the reply.
For the ofm , is that in seconds?
ofm and owd are in seconds.nanoseconds in the latest code.
Thanks for the reply gnn.
With the latest checkout version, I occasionally see our ofm as:
0.-0002000000
Should the minus sign be, before the 0? like -0.000200000
I think this maybe a bug
This is a bug which I have recently fixed in the ptpv2d branch.
Unfortunately it looks like the complete code base has initially simply been copied to create the v2 branch, and after this both code bases have been changed in different ways.
I'll try to discuss with the other developers how to proceed. IMO the best solution would be to change the code base so that utility functions like printing time stamps, or even the servo loop used to discipline the system time, would be available only in a single source file.
Just a quick note to say that we have talked via email about this and we are in agreement on creating a single source base going forwards. The initial goal was to have a working v1 and a working v2 before making the single source base. That goal is now nearly achieved. Once both daemons are stable we will create a new daemon, likely to be called v3, which supports both v1 and v2.
Yup, it's a bug. Can you enter a bug in the tracker for it?
Thanks,
George
Offset in ptp v2 - ID: 2997590
Bug filed and patched!
Good work!
Just a quick note: I've just committed a patch which fixes printing negative time offsets in ptpd v1, similar to the earlier fix in v2.
Thanks again.
So, I am assuming this is in trunk…