>Number:         11046
>Category:       misc
>Synopsis:       Need daily check of NTP daemon
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    misc-bug-people
>State:          open
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Tue Sep 19 19:55:01 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     Paul Sander
>Release:        <NetBSD-current source date>NetBSD 1.4.1
>Organization:
Paull Sander
	
>Environment:
All
	
System: NetBSD bugs 1.4 NetBSD 1.4 (GENERIC) #0: Sat May 8 20:54:59 PDT 1999 root2@c610:/usr/src/sys/arch/mac68k/compile/GENERIC mac68k


>Description:
	
	On some architectures (notably the Mac68k port) the clock is given a lower
	priority than other I/O, which causes causes dramatic shifts in the
	system time under heavy disk access such as that which happens when the
	/etc/daily script runs.  If the disk is sufficiently large, the shift
	can be over 15 minutes, and it can occur faster than the xntpd daemon
	can keep up.  As a result, the xntpd daemon kills itself and must be
	restarted.

	I propose adding a new check to the very end of the /etc/daily script
	to invoke this command:

		echo lassoc | ntpq localhost

	This should be chosen with the customary check_yesno call with a
	suitable argument.

	This will alert the sysadmin when it is necessary to restart the
	NTP daemon due to this condition.
>How-To-Repeat:
	
	Start xntpd on a system that has more than 2.0 GB disk and wait for
	/etc/daily to complete.
>Fix:
	
	Until the clock is fixed (there has been discussion of this in the
	mailing lists), see above.
>Release-Note:
>Audit-Trail:
>Unformatted: