Restart of my node is impossible

Support and Troubleshooting
1

Hi

Raspberry 4 with UmbrelOS 0.4.17
I’ve had my internet connection down for a few hours, and my node was unreachable when internet was up again (no more IP to ping).
I’ve unplugged (I know I shouldn’t have, but I had no alternative), and now I cannot start anymore the dashboard, my node is only accessible with ssh.
I’ve tried those 3 groups of commands with no luck:

sudo systemctl stop umbrel-startup.service && docker system prune --force --all && sudo systemctl start umbrel-startup.service

sudo docker kill $(sudo docker ps -aq) && sudo docker rm $(sudo docker ps -aq)

cd umbrel
sudo systemctl stop docker.service && sudo systemctl stop docker.socket && sudo rm -rf /var/lib/docker/* && sudo umount /var/lib/docker || true && sudo rm -rf /mnt/data/docker/* && sudo reboot

When I ran these commands, issue is still the same:
Instead of the login dashboard, I have this message:
Error: Failed to start containers
New to docker stuff, but not to linux, feel free to suggest advanced commands :wink:

2

READ and apply:

3

My node had a static IP for months.
And I don’t have any machine runing windows here for years now.
Is this answer supposed to be related whith my problem ?

4

That guide is not only for windows. Same apply to linux or mac.
When you said “no more IP to ping” means your router, changed the original IP dynamically assigned by the router and not established in your node machine and your PC you are using to access your node.

Static IP means, you literally edit a conf file in your node and set that IP that always will be. Your router then will just accept it and forward it.
Your local PC you use to access your node also is better to have it well established in local hosts file, where you indicate that umbrel.local name have x IP. In that moment whatever you do with your local PC, will always go to that local IP and will not searching around where is umbrel.local machine.

This procedure it makes go away a lot of trouble. I am trying to help you.

Now you start, running some commands, without waiting the necessary time for your node to catch up and “repair” itself. So you mess things up more than was necessary, now you have a double troubleshooting.

Failed to start containers error could be easily fixed with just relfashing the mSD card. Yes, that little fucker is the first one crashing at a sudden unplug the power cord situation.
So start by doing that. Stop your node, normally with shut down. Unplug the mSD card, put it into Balena etcher like you did firrst time and flash the latest version of umbrel.
Done, put it back and start.
Have patience! Let the node catching up.

5

When you said “no more IP to ping” means your router, changed the original IP dynamically assigned by the router
No. Static IPs are given on a per MAC address basis (in my home, I mean).

Now you start, running some commands, without waiting the necessary time for your node to catch up and “repair” itself. So you mess things up more than was necessary, now you have a double troubleshooting.
How do you know I haven’t wait enough time between the “groups of commands”?
It’s possible I haven’t, but for one of them, I let them run all night long, and message was still the same when I woke up.

Failed to start containers error could be easily fixed with just relfashing the mSD card.
I flashed another mSD (with dd like I did the first time), and no new device appeared in my local network after the boot.
I’m going to try again.

Thanks

6

No, that is not good. Is still managed by your router.
You should do it as described in the guide.

try looking into bitcoin and lnd logs
use this command:

cd umbrel
docker-compose logs bitcoin

# then run
docker-compose logs lnd

See if you can find something bad in there.

7

No, that is not good. Is still managed by your router.
Yes, my router gives the IP 192.168.x.x to the device connected which has the MAC address xx:xx:xx…

docker-compose logs bitcoin and docker-compose logs lnd return the same:
Attaching to
(and nothing else)

8

Then you have more serious problems there. I hope you have a backup of your channels.
If not, at least copy manually the file, now that you still have access to the drive.
make a copy for these files, just in case:
/umbrel/lnd/data/chain/bitcoin/mainnet/channel.backup
and
/umbrel/lnd/data/chain/bitcoin/mainnet/wallet.db

/umbrel/lnd/data/graph/mainnet/channel.db

9

I have a backup, and I’ve made a clone of my mSD before I flash it, later today.

Thanks

10

Node back online :slight_smile:

Thanks @DarthCoin for the help.

As I said, I had flashed another mSD, but I had no IP assigned to my node with this mSD. Now I know why: although it was written “16Gb” on it, my OS only recognized it as a 8Gb mSD, so it couldn’t boot.
I found why only by pluging in a screen and a keyboard, of course.

So I’ve reflashed the former mSD, and it booted fine :wink:

I had to reset lnd.conf file too, but everything seems to be ok now :sweat_smile:

Thanks again (I’ve learned a few things, cool !)

1 Like
11

Cool! I am glad you solve it!