the requirements for "tt"

Rick Jones raj at cup.hp.com
Sat Nov 6 03:19:43 UTC 1999


I'm trying to ascertain the "hard" requirements for the tt global. Just
how "closely" does it need to track the real timeofday. 

I gather from the frequency of use versus the number of calls to gettime
that the closeness to actual time is not absolutely critical. Most of
the time, it appears to be used to measure deltas in time for ttl
expirations and retransmissions. 

There are a couple places in db_load/db_unload/db_update where it
appears to be used to create a timestamp for a logging message. I'm
wondering just how close to "real time" that needs to be?

Basically, the upshot of all this is to see if gettime() can "safely"
(nice relative term I suspect) be modified to call something that at
least at first is the system's gettimeofday value, but could drift away
over time - for example because it does not get an adjtime
adjustment...it would however still have a monotonic increase.

rick jones
-- 
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to email, or post, but please do not do both...
my email address is raj in the cup.hp.com domain...


More information about the bind-workers mailing list