[AusNOG] Telstra outage?

[Mark Smith]

On Tue, 14 Jul 2026 at 19:16, Ross Wheeler <ausnog at rossw.net> wrote:

>
>
> On Tue, 14 Jul 2026, Mark Smith wrote:
>
> > So it seems a "nice to have" upstream NTP service that wasn't accurate
> > and people didn't really care about became a need to be accurate on a
> > network access device.
>
> This might sound conceited but I've never trusted "upstream" anythings
> because I don't know how (or IF) they give a damn, unfortunately often
> it's "we don't".
>
> If it's technically and financially PRACTICAL for me to do it in-house, I
> do. Always have.



I'm guessing the networking people who originally pointed this device at
the company's NTP service when they set it up didn't realise how critical
it might be to the device, and the upstream NTP service would have had been
reliable for some time otherwise the reboots would have occurred not all
that long after the device was installed (and there were 8 or more of them
around Australia IIRC that weren't rebooting.).

I think most people would believe that accurate (wall) clocks in networking
devices aren't quite a need to have, because it would appear that their
only use would be timestamping logging messages. All of the routing
protocols we use e.g. OSPF, BGP etc., while they might use time and timers,
don't use wall clock time.

I agree fundamentally, you can't assure what you can't control.

Regards,
Mark.




> Because *I* care about me and I know very few others do.
>
> R.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ausnog.net/pipermail/ausnog/attachments/20260714/ed03c86d/attachment.htm>