[Pool] Off by a second leap second failures

Martin Burnicki martin.burnicki at burnicki.net
Sun Jul 12 08:10:56 UTC 2015


Harlan Stenn wrote:
> John Winters writes:
>> Under Debian, ntp runs as its own user and group and I seem to have 
>> installed the file with:
>>
>> -rw-r----- 1 root root 10386 Apr 14 17:15 leap-seconds.3629404800
>>
>> Mea culpa.  Odd that ntpd didn't complain at start up about being unable 
>> to read one of its configuration files, and also reported that it was 
>> ready to leap, but I think that's probably the explanation.
> 
> Looking at the code, there was clearly a conscious decision to avoid
> always putting these messages in the logfile.  They are only logged if
> "logall" is provided.

IMO things like this should *always* be reported. That's what logs are
for. If things go wrong, or don't work as expected, as in this case,
then admins eventually don't notice this if there#s no log entry, and
then just wonder and start investigating why things don't work as expected.

This has been discussed when 4.2.6 was rolled out, and many log messages
which appeared by default in 4.2.4 weren't generated anymore by default.
There's a bugzilla issue related to this.

> I'm not sure if that's a bug or a feature, but it was a deliberate
> choice.

I know there are folks out there who prefer that there are as few as
possible log entries by default, just to prevent customers from asking
them what a log message means, but I think it's a bad idea not to log by
default when things go wrong.

Martin



More information about the pool mailing list