Sleepy UmbrelOS 1.3

Support and Troubleshooting
1

v1.22 worked fine & stable (!) for several weeks supporting Bitcoin node, Electrs, Lightning node, Seafile and Wireguard server running on miniPC 4GB/800GB. The Bitcoin node was synchronized many months ago.

Noticing the lively sign of UmbrelOS update v.1.3 and “In a full trust” in the new Umbrel team product, I updated and it went without any problem. All run fine and well with CPU load about 30-70%

In 4 days I noticed that CPU load was 3-5% and bitcoin stop synchronizing, the whole operation looked like Umbrel became sleepy.

The local access via the UmbrelOS web interface became very slow.

Had to reboot and it became lively again.

The story repeated in 7 days again. Reboot, fine and in 7 days again the same effect: Sleepy Umbrel, no syncronisation.

Following the resent post by 5tubb0rn, I added the “reindex=1” to the conf file, and after reboot began downloading the entire blockchain once again.

But it was good in 1.22!

Why releasing the buggy version?

1 Like
2

Well, let’s go step by step, hoping that it helps you.

  • Your context is that you run umbrel on a mini-pc with 4Gb ram and 800Gb disk.

  • Your umbrel runs Bitcoin node, Electrs, Lightning node, Seafile and Wireguard server.

  • You upgraded from v1.2.2 to v1.3 and the upgrade went well and without problems.

  • After about 4 days you noticed that bitcoin core was not syncing because there was no CPU load. You rebooted, it seemed to sync again and days later it went back to the same thing: bitcoin core was not syncing because there was no CPU load.

  • You followed a post by 5tubb0rn (I don’t know which post) and reindexed the blockchain and, according to you, the blockchain has started to download completely again.

If I’m missing something or misunderstood some point, please correct me.

First let’s clarify some points because I think you don’t know what you’ve done.

1.- When you reindex the blockchain (reindex=1) you don’t download the entire blockchain again. In reality, you’re ordering your bitcoin core to “check” the blockchain you already have for errors and “index it again”. This is to put it simply. You’re not downloading anything again. As long as you’ve only done that: reindex=1.

2.- The v1.3 has given many problems when updating. I don’t think that’s your case because you say that everything went well until days later when you noticed that bitcoin core didn’t sync. I really don’t think it was the update. At least I draw that conclusion with the information you’ve provided.

3.- And here we come to the interesting part. You say that everything runs on a mini-pc with 4Gb ram and 800Gb disk. The blockchain alone weighs 706 Gb at the time I’m writing this. Then add the files from umbrelos, Electrs, Lightning node, Seafile and Wireguard server. In 7 days your disk should have filled up and run out of space. Logically, there was nowhere to continue downloading the blockchain. Maybe if you look in your umbrel»settings, under storage… it might show you the status of your disk. It can also be done using commands via ssh but I don’t want to complicate your life.

In short, you followed a piece of advice that, I’m almost sure, had nothing to do with your problem.

Apparently there is a silver lining to all this, my friend. Maybe you just need to change your disk to a larger one.

3

Thanks for the detailed assessment and recommendations.
Some clarifications:
1/ Actual SSD=8000GB (sorry, I missed one 0 in a hurry)
2/ Searching how to improve the umbrel performance I found the comments

and

recommending to add the line
reindex=1
to the umbrel-bitcoin.conf
I did it in the terminal using nano, saved it and restarted umbrelOS.

It began resyncing as it was a new installation, now finishing about 90%
However now after checking the content of the file “/home/umbrel/umbrel/app-data/bitcoin/data/bitcoin/umbrel-bitcoin.conf” I found that it does not have the line reindex=1 in it (!)
Having this result I do no understand your your comment #1 Re reindex=1:
supposing that I had saved the update umbrel-config correctly it should not have began reindexing, but it did.
Unless I mistakenly had no saved correctly (I had not checked before restarting being sure …)
How you see this situation?

4

Well, based on everything you told me, I see that it has worked correctly. So you understand it better.

When you use the “reindex=1” command, it tells the node to reindex the blockchain from scratch. Immediately and automatically after starting to index, the node deletes the “reindex=1” command.

This is as a preventive measure in case you forget that you added said command. If it were left there, the node would re-index from scratch every time you restart the node, for example.

In fact, it is recommended to consult the .conf file where you added said command after the node starts indexing. This is to confirm that the command has been deleted or, failing that, delete it to avoid what I explained in the previous lines.

So far, everything seems correct. The node has correctly done what you asked it to do.

I have re-read from the beginning and I have some doubts.

You say: "…In 4 days I noticed that CPU load was 3-5% and bitcoin stopped synchronizing, the whole operation looked like Umbrel became sleepy.

The local access via the UmbrelOS web interface became very slow.

Had to reboot and it became lively again.

The story repeated in 7 days again. Reboot, fine and in 7 days again the same effect: Sleepy Umbrel, no syncronisation.."

Did you only notice that? For example, if the node had stopped synchronizing and you entered the node via the web… What did it tell you?.. Was it 100% synchronized?.. If it wasn’t synchronized it would show you what it needed to reach 100%.