[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ntp-hackers
Subject:    [ntp:hackers] GSOC: More logging requests
From:       Hal Murray <hmurray () megapathdsl ! net>
Date:       2009-06-25 0:16:46
Message-ID: 20090625001647.3B021BCEF () ip-64-139-1-69 ! sjc ! megapath ! net
[Download RAW message or body]


I'm talking about clockstats now rather than syslog.

I'm using the NMEA driver as an example because I'm familiar with it.  Other 
drivers probably have similar opportunities.

The NMEA driver gets a line of input every second.  Currently, it logs the 
next one after it gets polled which is normally every 64 seconds.  That 
mostly works out OK, but often I'd like more info.

Each line contains a good/bad flag.  With normal logging, I only get to see 
the last sample.  How many of the last 64 were good?

One option would be a flag to log everything.

Another option would be to put everything into shared memory where another 
program could grab whatever it wants.

Another option would be to count the good and bad lines and print those 
counts by appending them to the next line that gets logged.  But there won't 
be a next line if the cable has been unplugged so you have to handle the 
no-input case too.


NMEA has an interesting quirk in that you can setup most of them to give you 
more info than you need for just timekeeping.  The log-everything and/or 
shared memory could work well for that.  In this case, "everything" has two 
meanings: all types of NMEA sentances vs all of a particular type (rather 
than just one per poll interval).  So we probably want 2 flags.



-- 
These are my opinions, not necessarily my employer's.  I hate spam.



_______________________________________________
hackers mailing list
hackers@lists.ntp.org
https://lists.ntp.org/mailman/listinfo/hackers
[prev in list] [next in list] [prev in thread] [next in thread]