Issue with LND Node Failing to Start GRPC Services After Restart

[nemanja-z]

After restarting BTCPay Server, I noticed that my LND node fails to start the GRPC services.

Details:

• I executed the chantools compactdb command on my channel.db file, which ran without any issues.
• However, the size of the resulting compacted.db file is identical to the original channel.db file, which is 4.7GB.

This raises a few important questions:

1. Does the compactdb command always fail in the case of a corrupted channel.db?

2. Does the size of the compacted.db imply an issue with the channel.db file?

---

I'm curious to hear if others have encountered similar situations and whether there are additional steps or checks I should consider to resolve this issue. Any insights on whether the compaction process behaves differently with corrupted databases or if there's a way to further troubleshoot the GRPC service failure would be greatly appreciated.

[guggero]

fails to start the GRPC services.

Hard to say what actually happened without any logs...

Does the compactdb command always fail in the case of a corrupted channel.db?

Most likely yes.

2. Does the size of the compacted.db imply an issue with the channel.db file?

No.

I think you should check the logs for the actual error, then we can help you with more specific actions.

[nemanja-z]

The issue is basically the same as this one: lightningnetwork/lnd#6179. What I forgot to mention is that this LND instance used to take a few minutes to start in the past, but it always did. What else can I provide? The log file shows the same messages as the screenshot in the linked issue.

[guggero]

Are you sure there are no errors? Can you check the container log of the lnd container? I think certain panics are only printed to standard out and don't appear in the on-disk log file.

[nemanja-z]

Compacting the database did the trick. It took more than 20 minutes for the node to start. Regarding the compacted database size, I initially enabled auto-compaction in BTCPay Server and then used chantool to compact the database in a separate folder. The channel.db was already compacted, which might explain why the sizes of channel.db and compacted.db are the same. Unfortunately, I didn’t check the size of channel.db before the compaction. Thank you for your help.