Actually that only shows the signature of the file. It doesn’t verify it.
The OpenWRT trust management is based on the gpg signature of the sha256 file. Once that is verified you can use the sha256 data in that file to verify the actual firmware file.
Ok, I did it by just manually downloading this file https://git.openwrt.org/?p=keyring.git;a=blob_plain;f=gpg/88CA59E8.asc and then doing gpg --import ~/Downloads/gpg_88CA59E8.asc For some reason the key at keys.openpgp.org is different than the one at git.openwrt.org
Edit: After some digging, it looks like keys.openpgp.org strips down identity information from public keys if an email is not verified, and gpg doesn't allow importing such keys.
Like the error message says, you need to use the "factory image" instead of the sysupgrade image. You should be able to flash it by selecting the "force upgrade" there. And unselect the "keep settings".
The underlying reason is explained in that error message that you see: Linksys had defined too small partition for the new kernel versions (used in 21.02 and master) for WRT1900ACv1 (mamba) and WRT32X (venom). So, the kernel partition has been enlarged for those two, but that prevents using the sysupgrade, and you need to use the "factory" image to start fresh.
https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=8458ebe18b84702039020c523018e39f564ccf6c;hp=7939d4a1b13b0c44dbfa4cd03180fe4cd5b2fe6b
At the same time you also get the switch config change from swconfig to the new DSA, so you should not restore your old config from backup...
I was reading some other post about WRT32X with too small partition but I did not see the WRT1900ACv1 one. That makes sense. Thank you so much for your help.
Congrats on this big release. Sticking with occational master snapshot updates on my WRT32X for now since it's stable and wifi is good with it. Great job to the devs thank you.
Yeah, strangely in addition to the oldest device in the series (WRT1900ACv1) Linksys again caused the same problem to the newest device (WRT32x), although they had already used a larger kernel space for all the devices released in between, including the WRT3200ACM, the older twin sister of WRT32x.
I wouldn't keep the settings, for sure. It is serial recovery if you forget.
I usually keep the old setting when upgrading but in this case I guess it won't hurt setting up my routers as new again. I have some old backups so I just need to loop through them if needed.
Yes, that fixes the problem.
v1.5.0-2
v1.5.1-1
It can't even boost to full bandwith. Btw im ussing your script, that could be the problem too.
nice... I will check my end...
Just like RC4, this version has issues with Linksys WRT1900ACS v2. Clients loose connection to internet while connected to wifi. Can be fixed (temporarily) if you disconnect and reconnect wifi.
roughly how long does it take, before the clients loose Wifi again?
Varies, but minutes. Doesn't seem like a lease renewal issue.
in LuCi advanced wifi settings of each radio there are 2 options:
Disable Inactivity Polling, default off
Disassociate On Low Acknowledgement, default on
Maybe try inverting these separately for a quick test, to see if it makes any difference.
You could also try current master image as well for a comparison - at least Wrt32x and Wrt3200 had ongoing issues and users claim the master works a lot better.
Congrats to everyone involved in the development of this new stable version. It makes the already best even better.
For a 88W8864 device check post as to resolution.
What are the differences between rc4 and stable?
Read above, a few messages ago...
Would you consider opening an issue upstream? If there are regressions, they should know about it.
