ntpd leapseconds file didn't update automatically

rwv37

rwv37

I noticed that I've been getting messages once a day saying that the leapseconds file had expired for about a week now. I checked the file, and it indeed listed an expiration date in the past.

I did a service ntpd fetch, and the file was updated (I may have also done a needfetch? not sure). The file was updated. So that's cool, but... wouldn't it normally have been automatically updated somehow? Why might it not be (or at least not have been) automatic on my machine?

Here's the only rc stuff I have related to ntpd:

Code: 
ntpd_enable="YES"
ntpd_sync_on_start="YES"

Thanks in advance.
 
See daily/480.leapfile-ntpd. It should have worked.... may be anticongestion (defined in /etc/defaults/periodic.conf) has something to do with it?
 
Maybe it's automatically checked 30 days prior to expiration only.
/etc/defaults/rc.conf has an entry below.
Code: 
ntp_leapfile_expiry_days=30     # Check for new leapfile 30 days prior to
                                # expiry.
 
Thanks, both of you.

bakul: The source for anticongestion() looks to me like in the case where it does anything at all, all it does is cause a sleep for a random amount of time up to an hour; doesn't look to me like it will cause the intended periodic jobs to actually be skipped or fail.

T-Aoki: Looks to me (in /etc/rc.d/ntpd) that the ntp_leapfile_expiry_days will cause the fetch to be skipped if and only if periodic is being run more than that number of days prior to the expiration (as opposed to not exactly that number of days prior to the expiration). I could imagine that maybe the machine happened to be completely down for that entire particular day, but it's very difficult for me to imagine that it was down for a whole month without me recalling anything of the sort.

Uhhhh... maybe I somehow turned periodic off entirely? Is there a way to do that? How can I check if I did?
 
Are you running something quite heavy-loaded while the periodic check for leapseconds attempts to run every day?
In such a case, periodic script could fail to start.
 
T-Aoki I guess it's possible, depending on exactly how heavy it would need to be, but I don't know what time that would be. I'm afraid I don't know much about periodic. Would it (or something) log its failures to start scripts? Or is there a way to tell when they are supposed to run?
 
I don't have the answer, but definitely on some of my FreeBSD servers I seem to have to be more proactive in service ntpd fetch and service ntpd restart to make the warning go away. And other servers just seem to do it right themselves. But I've not worked out why yet; but there's definitely something going on.
 
You must log in or register to reply here.
Back
Top