Do you have some more details about how you fixed this? 
I just set optional: true for netplan for fix this
апр 15 14:36:46 gateway systemd[1]: Starting Wait for Network to be Configured...
апр 15 14:38:46 gateway systemd-networkd-wait-online[495]: Event loop failed: Connection timed out
апр 15 14:38:46 gateway systemd[1]: systemd-networkd-wait-online.service: Main process exited, code=exited, status=1/FAILURE
апр 15 14:38:46 gateway systemd[1]: systemd-networkd-wait-online.service: Failed with result 'exit-code'.
апр 15 14:38:46 gateway systemd[1]: Failed to start Wait for Network to be Configured.
1 Like
Done.
2 Likes
I cannot create network bridge for lxd
I got
"Error: failed to create local member "lxdbr0" in project "default": Failed to run: ip link add lxdbr0 type bridge: Error: Unknown device type.
I think you need to enable Canonical Kernel which has all necessary modules enabled
Update:
Ubuntu Server 22.04 LTS Kernel 5.17.5-generic
Changed
- Kernel 5.17.5
- Kernel based on original Ubuntu server build config 5.15.0 – generic. (Included all modules and boot time increased by 2 times. From 30 sec to 78 sec on a class 10 sd card.)
- ESP & boot partitions reduced.
- removed cloud-initramfs-copymods
SSH login\password
user\star5
Also you can install kernel severally
I think in the future we can find good set of modules for avoid long time of boot. Generic kernel is redundant functions.
3 Likes
Thank you a lot,
Generic is good for everyone because they can use it however they want without need to rebuild the kernel
1 Like
From your log:
No space left on device
Need more space on disk. About 3 GB
1 Like
Thank you for this! I just started tinkering with my VisionFive, but I have absolutely no patience for RedHat. Everything’s years out of date if it’s even packaged at all, the config files are in weird places and have bizarre defaults, etc. ad furiatum.
@jershell I notice something wired and I don’t know if it VisionFive limitation or Ubuntu 22.04
I cannot get IPv6 from the router and I tried everything from edit netplan to ifconfig command
If you’re referring to RH in general, that’s just the speed they move. Fedora is faster tracking, but that’s just not the market they serve.
If you’re referring to the RISC-V build specifically, it’s based on F33. They even have a nightly build of this 2020 release that tracks updated versions of criticial packages. It’s indeed odd that we’re piling onto an unreleased OS/arch combo that’s well past EOL.
In a post somewhere, they said they’ve kept that deeply frozen to avoid having too many moving pieces. When Silicon, specs, compilers, kernels, and all the packages are all changing weekly, debugging can be tough. So they’ve provided a long-term build to emulate installed base. That helps keep everyone on an agreed-upon ABI, for example. I read somewhere that Fedora group was working on rebasing to a contemporary version, but publicly visible progress on the repo seems to have halted this year.
I’ve had success with the official images from GitHub - starfive-tech/Fedora_on_StarFive. Needing to manually un-zstd the images before feeding them to Balena is unfortunate.
I don’t know if more coding volunteers would help. It’s not clear what the issue is. There are details that don’t inspire confidence.
By default Ubuntu setups a link-local ipv6 simular fe80::bla::bla::bla
Can you use the linklocal address for a test?
For do it
From your device run ping6 link-local-ipv6-of-your-router%device-name
Example
ping6 F80:000218:e7ff:fe16:fb97%wlan0
If it will work it means ipv6 working.
And need looks to protocol that u use for configuration for accepting ipv6. Radvd or dhcp6 or another.
I tested ipv6 a link local and had success. I can test client, but for this i should know which a client.
I tried many ways, first set static ip using netplan from my router and that work for ping vf and I cannot use to ping ipv6.google.com for example
then I tried IPv6 from route48.org and WireGuard without success
Okay. I checked it and it working.
I used a network-manager.
- Create a file wg0.conf. For example vi /home/user/wg0.conf and paste wireguard config from route48.
After import to network-manager by command
nmcli connection import type wireguard file ./wg0.conf - Allow use wireguard type of connection for network-manager. By defaut its not allowed. You can see it by the command
nmcli dev status
wg0 wireguard unmanaged --
For fix it neccessery change configuration of network-manger by edit/etc/NetworkManager/NetworkManager.conf
on this
[main]
plugins=ifupdown,keyfile
[ifupdown]
managed=true
[keyfile]
unmanaged-devices=*,except:type:wifi,except:type:wireguard
[device]
wifi.scan-rand-mac-address=no
- Restart network-manager service.
sudo systemctl restart NetworkManager - NetworkManager up tunnel automaticaly. Reconnect it by commands
nmcli connection down wg0andnmcli connection up wg0 - Check it.
nmcli dev status
and
wg show
Look at this moment
transfer: 10.63 KiB received, 9.46 KiB sent
Some packanges transfered and receivered.
now we can runping6 ipv6.google.comorcurl -6 https://ipv6.icanhazip.com
Also u can disable an autoconnect
nmcli connection modify wg0 connection.autoconnect 0
ok, I will try it connect from NetworkManager ( I tried wg-quick )
second question if possible
Can I know why VF didn’t get IPv6 address from my router like other devices
I dont know. Configuration depend on protocol. What you use a protocol radvd or dhcp6?
I think my router use SLAAC
BTW, How do I make the newest kernel for myself? Can you make a detailed tutorial for us in your spare time? or create a scripts repository on GitHub.
appreciate sincerely.
