Re: [chrony-users] Two Issues

[ Thread Index | Date Index | More Archives ]

On 03/12/2015 10:16 AM, Miroslav Lichvar wrote:

On Thu, Mar 12, 2015 at 09:58:34AM -0400, Weary1 wrote:

I've been using the fully-qualified host names for two main reasons:

1. - Those servers (and the networks they're hung off of) are not under my
control.  So what happens if/when their IP address assignments change?

Yeah, that would be a problem, but I'd not expect this to happen
frequently with public NTP servers.

Normally, I would assume the same thing. But OTOH, I *have* had some frustrating issues getting chrony to use my "#1 First-Choice" Tier 1 server (at, which is geographically close to me). Apparently, for a fair while, their published hostnames and IP addresses were NOT right. I've forgotten now how I finally got it all straightened out.

2. - I would prefer to see the server names in the listing produced by a
"chronyc sources", as it's easier to recognize which is which.

It would still print the names, chronyd works only with IP addresses
once the names are resolved and chronyc translates them back to names
(unless the -n option is used).

OK, then I guess that obviates one of my reasons to use the hostnames in "chrony.conf". Good to know. Thanks.

BTW... This is far from important, but I'm curious: Sometimes, but ONLY sometimes, when I issue a "chronyc sources -v" command, the response hesitates significantly (up to several seconds) at one or more points in the server list, as if it is having trouble finding the data. What would cause this?

Half-baked idea #647:  I doubt this is "kosher"; but would it foul things up
too badly if I were to define each source "server" twice, once by host name
and once by raw IP?

I'd not recommend that.

Yeah...  That idea was something of a longshot from the get go.

Delaying start of chronyd until network is probably a better idea.

Any ideas on exactly how to do this? I've yet to really get my head wrapped around the Debian start-up sequence, with its branched "run levels" and funky coded script-filenames scheme, and all that; and I'm more than a little gun-shy about making changes I'm not sure of, lest I *really* screw something up.

You could also try cleaning /etc/resolv.conf before chronyd is
started, so it would keep trying resolving until the DHCP client
starts and rewrites the file.

Again, how exactly would I "clean" /etc/resolv.conf ?

It already looks pretty clean to me; just four lines:

    domain <my domain name>
    search <my domain name>
    nameserver <IP address of my LAN's FW/router)
               (which serves DHCP to the rest of the network)
    nameserver <IP address of my ISP's primary DNS server>

The best option here would be update to a more recent version.

Well, in the case of the workstation, the (K)ubuntu crowd is already annoyingly "update-happy" by my yardstick (a day rarely goes by without "Yet Another Update" to deal with; even new kernels are almost a weekly occurrence). So I suspect something newer will come along sooner than later; but I have no specific knowledge on this.

As for the server, that will eventually happen automagically when what is now "Testing" (code name "Jessie") becomes Debian 8.0 Stable. That is rumored for April; but if history is any guide, May is a better bet.

In the meantime, I suppose I *could* Apt-Pin chrony to use the version now in "Testing" (1.30-1). But I strongly prefer to avoid such "kludges" whenever possible, as they tend to make things messy down the road.



To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject. For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.

Mail converted by MHonArc 2.6.19+