@Tsahi_Etziony @Dmitry_Krupnik
Just some feedback for you. We performed a CPUSE Blink + JHT clean install last night on a secondary gateway. The blink package went nice and quick, and the device came back up with all interfaces configured with the correct IP addresses. We simply needed to re-establish SIC, push policy and we thought we were good to go. As we started poking around to verify everything, we released the config did not get pulled over entirely, it appeared hit and miss. We were missing NTP time server, DNS server (which caused contract lookup, AV/TP update lookup and HTTP Categorization lookup errors) OSPF routing config completely missing, the description on some of the interfaces was missing (some had them though which was interesting) and other various little missing configs. We decide to re-import the config back in as we were playing whack-a-mole with what was missing. As soon as we re-imported the config back on, bounced the box, pushed policy, bounced the box again, then everything began working as expected.
I'm not sure if my experience here was a one-off, but I figured it would be good to communicate this back just in case it represented a larger problem. I will be performing another Blink upgrade in the exact same fashion on a different cluster tomorrow night, and then we will be upgrading the other cluster nodes some time next week. I'll report back my results of those.
Also, the Blink + JHT Gaia 3.10 did get removed from my package list on the 4000 series gateways.