Request 230 was acted upon.
This is a comment. It is not sent to the Requestor(s).
_________________________________________________________________________
URL: https://rt.openpkg.org/id/230
Ticket: [OpenPKG #230]
Subject: Script rc.ntp is likely flawed in section %hourly
Requestors: [EMAIL PROTECTED]
Queue: openpkg
Owner: thl
Status: open
Transaction: Comments added by thl
Time: Thu Aug 07 13:42:17 2003
_________________________________________________________________________
On Wed, Aug 06, 2003, Ralf S. Engelschall wrote:
> On Wed, Aug 06, 2003, Michael Schloh v. Bennewitz via RT wrote:
>
>
>
>
>
> > # /cw/bin/ntpdate `grep "^server" /usr/opkg/etc/ntp/ntp.conf | awk '{ printf("
> > %s", $2); }'`
> >
> > 6 Aug 16:29:16 ntpdate[21868]: the NTP socket is in use, exiting
> >
> >
> >
> >
> >
> > This message is not diagnosable through normal means of testing ('rc hourly ntp'
> > produces no error text).
> >
>
>
>
> This is nothing special if ntpd is already running. ntpdate
>
>
> can only adjust the time if ntpd is not already running.
>
>
>
>
>
I didn't know that. This renders some combinations of ntp_ostart,
ntp_daemon and ntp_hourly useless and i agree they should be integrated
into a single "mode" option. I'll use your patch as a base but will
change ntp_mode to be understood as one of
once = run ntpdate once at startup
daemon = launch ntpd
monthly = cron
weekly = cron
daily = cron
hourly = cron
quarterly = cron
All modes will run ntpdate once at startup (before the daemon is
launched). The less frequent cron options will satisfy dialup users.
The second issue was the usage of rcService for comparing variables
outside the %scope. I was not aware of this intended rcService
limitation. It was prooved to work in run-time testing but i will
change this in all affected packages, as this is not limited to the ntp
package. PR#232
______________________________________________________________________
The OpenPKG Project www.openpkg.org
Bug Database Interface www.openpkg.org/bugdb
Bug Database List [EMAIL PROTECTED]
