Flubster

As far as I can tell, 7.3.2 was released first and this is the catch-up security release.

Yeah, i consoled into the container and replaced it that way. SIO doesn't seem to update his containers once released therefore no real risk of it reverting.

replace piawg.py in the datadir to this: https://raw.githubusercontent.com/Richardarkless/pia-wg-config-generator/refs/heads/patch-1/piawg.py Reference: https://github.com/SpaceinvaderOne/pia-wg-config-generator/issues/2

7.3.2 didn't fix it for me, A310 Sparkle works on Arch VM 7.2.7, not on 7.3.2. Even though qemu was bumped. Do have migration set as well. Its listed in lscpi as I would expect, but just doesn't work. No obvious errors in syslog. So back to the 7.2 branch for now.

Other option as a thought - if you use something like cloudflare (or another ddns provider), you could set the unifi to update the ddns entry for a hostname - as long as WAN1 is static you can nslookup the hostname and if its your static IP you'll know you are on WAN1, should this change then you'll know you are on WAN2 or failover

How technical are you? Last time I looked (I decided it wasn't worth the effort) UniFi uses dpinger to check the interface and start the failover. You can create your own dpinger instance on the unifi hardware, which can run a script (on the unifi appliance) which in turn used ssh (and sshpass - be careful bad mmkay) to launch a script on the unraid box to "docker stop" containers. This should help. Its a pain tbh,

I would try this - but it does concern me that you've locked the source behind a private repo. Not saying you don't have your reasonings, but not being able to see the code is a major red flag.

yes - unless you bind the device to vfio-pci so unraid won't see it and pass it through to a linux VM with a newer kernel.

I suspect this is due to you putting the forums under cloudflare check but using an RSS/XML feed is failing on 403. Example: https://validator.w3.org/feed/check.cgi?url=https%3A%2F%2Fforums.unraid.net%2Fforums%2Fforum%2F7-announcements.xml Yet hitting the URL directly works Any chance someone can make some changes to the CF ruleset to allow rss/xml calls without CF attack mode interference please.

if this is the bug I think it is then you need to add the card to a windows machine, install the intel drivers and let it update the firmware. Its a bug.

Does the power button do a hard power cut, or does it do a soft shutdown? If the latter you can do this: Works a treat. Make sure you realise what this means (requiring ssh / console access for reboot / shutdown -h now) (I'm still on 6.12.15 so haven't tested on 7)

no response to the ticket. I haven't tried 11.6.0, which the latest image is on which may have actually fixed it.

Obviously the latest version of crashplan is 11.5.0 so this shouldn't be changed on the main repo, that should be the latest version. This maybe intentional by crashplan or could just be a bad release. Every other release seems to nag when it's out of date with crashplan attempting to auto update (and failing as it should on the container) every day. 11.4.0 doesn't. I find this suspicious. But I have raised a support request.

Haven't tried downloads, I will later when it's finished catching up.

I will report to crashplan, but forking the repository so it has the latest base image updates, but changing the dockerfile to pull 11.4.0.503 rather than 11.5.0.445, my uploads are back at normal speed. Interesting that crashplan doesn't moan it wants an update either.... So I'm running with it like that at the moment.