Bitcoin Forum

In a multi-wallet file configuration, I executed "reconsiderblock".
The console said "Executing command using ... wallet", making it appear only one is affected.
I chose another wallet file and executed the same command again.

Despite of what it says, does it only affect the single selected wallet or all of them?

If just one, what are the implications?


Background:

This is on v0.18.

Core crashed during catching up due to being out of space for the UTXO db. I fixed it and restarted.
It recovered and continued syncing blocks for about a minute, then stalled doing nothing.
The log revealed repeated "ERROR: AcceptBlockHeader: block ... is marked invalid" for the same block.
I restarted, the same situation continued, so I executed "reconsiderblock".
It seemed to resume okay.

The block status (valid or not) shouldn't be tied to a specific wallet, no. I don't see why it should.

The implication of the command is that it marks a block as 'to be reconsidered', e.g. if it was marked invalid through violating a consensus rule. It will thus be checked again.

It's possible that a block was corrupted due to unexpected exit of Bitcoin Core. Does it run normally now? Or does the issue persist?

The message will appear regardless if the command used is a wallet rpc or not, if you're using the GUI.
That's just a note to mention which wallet is selected in the drop-down menu above it.
But will only appear once until another wallet is selected.

If "(none)" is selected, it will say "Executing command without any wallet" instead.

Reference code: github.com/bitcoin/bitcoin/blob/master/src/qt/rpcconsole.cpp#L1031-L1038 (https://github.com/bitcoin/bitcoin/blob/73b61717a977fc9d23f1bae3f8620641a9dee1f3/src/qt/rpcconsole.cpp#L1031-L1038)

---

To see if a command is actually wallet-specific, enter help in the console;
Those under == Wallet == are the commands that affects only the selected wallet.

Thanks.

While it quit automatically due to the abnormal condition, it's still a graceful exit as evident by having everything logged until the end. Yet it always happens when out of space, so evidently the cleanup isn't good.

But it was on an older version. Maybe fixed in newer ones.

After "reconsiderblock" it's fine as far as I can tell, but not without it.


So it's sort of a bug or anti-feature in the case of non-wallet commands.

After a cursory look, it seems it may be possible to decide whether to show that "current wallet" console message based on CRPCCommand.category.

I think it's a way to show that the command executed in the console is equivalent to bitcoin-cli with --rpc-wallet arg.
Because in bitcoin-cli, you can also add rpc-wallet regardless if the command is a wallet rpc or not.

I did quick search and few possible error/corruption due to out of storage space isn't fixed yet, https://github.com/bitcoin/bitcoin/issues/26112 (https://github.com/bitcoin/bitcoin/issues/26112).


I'd guess it's easier/faster to code function which show such message on all command.

Easier, but misleading.

Interesting that the reported out of space condition there is also on writing to the chainset.
And also there reconsiderblock is reported as being good enough.