[ntp:hackers] Clues on the rackety problem

David L. Mills mills at udel.edu
Fri Feb 18 11:42:16 PST 2005


Danny,

As I said earlier, the smoking gun might be related to the fact that the 
problem occurs only when the radio is plugged in before the daemon is 
started. I would it could be fixed by looking at the length of the chunk 
returned and ignore if zero length. In other words, just clear whatever 
bit is involved and go back to sleep. I can't do that in the driver, 
which just returns without doing anything.

Dave

mayer at gis.net wrote:

>Dave,
>
>Any thoughts on the refclock failures like on rackety and Steve's
>CHU refclock?
>
>Danny
>----- Original Message Follows -----
>  
>
>>Danny,
>>
>>All the builds on all the machines here are on the machine A.whatever 
>>and the shared directory on pogo.
>>
>>Dave
>>
>>mayer at gis.net wrote:
>>
>>    
>>
>>>Dave,
>>>
>>>I've seen a number of problems with clocks, including rackety, cipo's
>>>and Steve's. I don't know what's going on and it's possible that it's
>>>related to the changes that I made to the input_handler() code.
>>>
>>>I looked at barnstable but I don't know where you do your builds on
>>>barnstable so I couldn't tell if something was wrong. 
>>>
>>>Danny
>>>----- Original Message Follows -----
>>>
>>>      
>>>
>>>>Guys (Danny?),
>>>>
>>>>If I unplug the radio on rackety, start ntpd and plug it back in, it
>>>>comes to life. If it is plugged in when started, the driver loops on
>>>>        
>>>>
>>a >>zero-length line. I suspect that for whatever reason the
>>particular  >>serial port on rackety generates a spurious interrupt on
>>    
>>
>>>>initialization  if the carrier lead is up or something like that.
>>>>        
>>>>
>>I'm >>antsy to fix this,  as rackety is our public time server of
>>record and >>it has lots of hungry  clients.
>>    
>>
>>>>Dave
>>>>
>>>>        
>>>>
>>_______________________________________________
>>hackers mailing list
>>hackers at support.ntp.org
>>https://support.ntp.org/mailman/listinfo/hackers
>>    
>>




More information about the hackers mailing list