home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!olivea!sgigate!sgi!rhyolite!vjs
- From: vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
- Newsgroups: comp.sys.sgi
- Subject: Re: Timeslave bug?
- Message-ID: <ufs8tf4@rhyolite.wpd.sgi.com>
- Date: 5 Jan 93 05:25:14 GMT
- References: <1993Jan4.140318.29296@erenj.com> <uf8kpe8@rhyolite.wpd.sgi.com> <matt-040193214031@wardmac2.med.yale.edu>
- Organization: Silicon Graphics, Inc. Mountain View, CA
- Lines: 30
-
- In article <matt-040193214031@wardmac2.med.yale.edu>, matt@wardsgi.med.yale.edu (Matt Healy) writes:
- > I have a related question. About every other month, a problem
- > at Yale Computer Center causes the campus timeserver to
- > jump around wildly. The usual result is a bunch of
- > "time changed..." messages in my SYSLOG until eventually
- > cron begins to complain.
- >
- > Is there a way of telling timeslave to perform a sanity check and
- > not follow such chaos? Maybe it could make *small* corrections
- > but ignore corrections larger than some limit except for noting
- > them in SYSLOG?
-
-
- Not reliably without changing the source.
-
- Timeslave assumes large and persistent changes mean the time standard
- has decided the date needs to be changed. If the master changes the
- time, and then holds the new value for a while, the slave accepts
- the change. How long is "a while" depends on how well the slave
- was following before.
-
- It might be effective to wildly increase the "-r" parameter.
- Turning on debugging with `killall -v -USR1 timeslave` or starting
- it with -d might be illuminating.
-
- Starting with IRIX 4.0 or maybe before, we fixed some bugs in
- cron's response to time leaping forward or back.
-
-
- Vernon Schryver, vjs@sgi.com
-