Cobia: What happened to ntp?

[Perry The Cynic]

I just upgrade a Mini-R test system from TrueNAS-SCALE-22.12.4.2 to 23.10 Cobia. It seems to have gone reasonably well, but I noticed that ntp seems to have disappeared: the ntp package is no longer installed, there are no ntpd or ntpq binaries in the $PATH, and the /etc/ntp.conf file is gone. The general systems setting for "NTP Servers" is still there, but it seems to have nowhere to go.

Is that a bug, or has Cobia switched to a super-secret time management approach I can't see?

Cheers
-- perry

 

[anodos]

I just upgrade a Mini-R test system from TrueNAS-SCALE-22.12.4.2 to 23.10 Cobia. It seems to have gone reasonably well, but I noticed that ntp seems to have disappeared: the ntp package is no longer installed, there are no ntpd or ntpq binaries in the $PATH, and the /etc/ntp.conf file is gone. The general systems setting for "NTP Servers" is still there, but it seems to have nowhere to go.

Is that a bug, or has Cobia switched to a super-secret time management approach I can't see?

Cheers
-- perry

chronyd

 

[Patrick M. Hausen]

chronyd

Why?

 

[anodos]

Why?

Don't worry, Core is keeping ntpd. ;)

 

[Perry The Cynic]

Ah, thanks. An entry in the release notes might have been nice.

Is there any way to remote-monitor chronyd with ntpq? Or is there any (widely available) tooling that can monitor both ntpd- and chronyd-based systems?

Cheers
-- perry

 

[winnielinnie]

Don't worry, Core is keeping ntpd. ;)

But that doesn't answer "Why?" What are you, a politician?

 

[anodos]

Ah, thanks. An entry in the release notes might have been nice.

Is there any way to remote-monitor chronyd with ntpq? Or is there any (widely available) tooling that can monitor both ntpd- and chronyd-based systems?

Cheers
-- perry

If you're concerned with specifically TrueNAS SCALE servers, we have a private API endpoint system.ntpserver.peers that we use for internal purposes for alerting on clock slew (this was added primarily for alerting on slew between cluster nodes). You can query it via shell middleware client midclt call system.ntpserver.peers. Since it has JSON output you can use jq to parse.

 

[jose2azulos]

If you're concerned with specifically TrueNAS SCALE servers, we have a private API endpoint system.ntpserver.peers that we use for internal purposes for alerting on clock slew (this was added primarily for alerting on slew between cluster nodes). You can query it via shell middleware client midclt call system.ntpserver.peers. Since it has JSON output you can use jq to parse.

That `midclt` call works on both Bluefin and Cobia, right?