For anyone that is interested, Juniper provided me with a solution which is genius if still a little hacky.
They suggest: 1. Deactivate existing NTP configuration 2. Set date back ~10 years root> set date 200403250000.00 3. Disable sw -> hw time sync (incl. at boot time via rc script) root% sysctl -w machdep.disable_rtc_set=1 root% touch /cf/etc/rc.custom root% chmod +x /cf/etc/rc.custom root% echo "sysctl -w machdep.disable_rtc_set=1" > /cf/etc/rc.custom root% cat /cf/etc/rc.custom 4. Re-activate NTP configuration 5. Reboot (doesnt seem strictly necessary, but maybe worthwhile as a test) So basically youre setting the hw clock back ~10 years which allows the FPC to come online. You disable sw -> hw time sync so even when running NTP, if the device reboots the hw clock is still in the past, the FPC will come online because the certificate is still valid, and then NTP will update the time on the box to the present. Hope that helps some one else out there. On 28 March 2014 18:21, Edward Dore <[email protected]> wrote: > Juniper have just published TSB16366 > (http://kb.juniper.net/InfoCenter/index?page=content&id=TSB16366) with a new > certificate and instructions on what to do about renewing expired licenses. > > Edward Dore > Freethought Internet > > On 28 Mar 2014, at 16:57, Tom Storey <[email protected]> wrote: > > For the newer J series its easy enough to load a newer JunOS image, > and in such a case sure, self inflicted. > > But the original J series stop at around JunOS 9.3 or so IIRC which > doesnt help. And sure, they may be well beyond EOL by now, but they > still have their uses, and Im sure many people would rather they just > kept on working. > > On 28 March 2014 13:16, Paul Mansfield <[email protected]> wrote: > > "So kit that has been EOL for 6 years, running code that has been > unsupported for 4 years has a bug that is easily worked around. > Whilst EOL was 6 years ago the hardware has some years support left to > run on it and the software can be easily upgraded to versions with > full support. Seems like a storm in a teacup to me and entirely > self-inflicted by the customer running truly ancient code for no good > reason." > > >
