You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
bitcoind rpc gets stuck in an infinite loop if incorrect password is provided #587
Yeah, as mentioned elsewhere this will only happen on regtest, as on mainnet we'd abort early if we're unable to update fee rate estimates. We also have to keep retrying to make progress there, as it's part of initialization of the chain syncing task.
We could however consider to make the delay larger, and we should see whether we can provide a more conclusive error message (although I'm not sure whether bitcoind RPC provides an 'invalid password' return code that we could lean on here).
the whole LDK-node process seems to get stuck.
Could you expand on what 'seems to get stuck' means?
@tnull can we open this issue? it's still a problem in v0.6.2.
Yes, the fix has not been backported to the 0.6 branch, so will likely only land with 0.7.0. Do you mind testing with main to see if the patch changes anything for you?
I'm unable to exit the program without suspending it and using kill -9 %1
That is still puzzling to me why the syncing loop would have any influence on signal handling. I really don't want to get into handing signals on our end as it shouldn't be the concern of a library that doesn't know anything about the OS/context/binary it's being used in.
Activity
tnull commentedon Jun 25, 2025
Yeah, as mentioned elsewhere this will only happen on regtest, as on mainnet we'd abort early if we're unable to update fee rate estimates. We also have to keep retrying to make progress there, as it's part of initialization of the chain syncing task.
We could however consider to make the delay larger, and we should see whether we can provide a more conclusive error message (although I'm not sure whether bitcoind RPC provides an 'invalid password' return code that we could lean on here).
Could you expand on what 'seems to get stuck' means?
tnull commentedon Jun 25, 2025
@tosynthegeek indicated interest in working on this.
tosynthegeek commentedon Jun 25, 2025
Yeah sure. Would love to pick this up
rolznz commentedon Jun 25, 2025
It's somehow stopping me from being able to exit the process. I have to force-kill it. ctrl+c which normally works fine in our app, just does nothing.
tosynthegeek commentedon Jun 26, 2025
@tnull please check #588
Add exponential backoff for sync failures
Add exponential backoff for sync failures
rolznz commentedon Aug 29, 2025
@tnull can we open this issue? it's still a problem in v0.6.2.
I'm unable to exit the program without suspending it and using
kill -9 %1tnull commentedon Sep 1, 2025
Yes, the fix has not been backported to the 0.6 branch, so will likely only land with 0.7.0. Do you mind testing with
mainto see if the patch changes anything for you?That is still puzzling to me why the syncing loop would have any influence on signal handling. I really don't want to get into handing signals on our end as it shouldn't be the concern of a library that doesn't know anything about the OS/context/binary it's being used in.