Zap wallet txns in pruned mode?

[-ck]

Since -zapwallettxes implies -rescan, it means its use is effectively mutually exclusive of using pruned mode. Does anyone have any suggestions for a workaround without redownloading the whole blockchain again?

[achow101]

Right click the transaction and choose "Abandon transaction" or use the abandontransaction command in the rpc console.

Or you can use something like pywallet and remove the transaction from your wallet file manually.

[-ck]

Right click the transaction and choose "Abandon transaction" or use the abandontransaction command in the rpc console.

Or you can use something like pywallet and remove the transaction from your wallet file manually.

Abandon transaction is greyed out so presumably it's subject to the same limitations as zap.

[achow101]

Abandon transaction is greyed out so presumably it's subject to the same limitations as zap.

No, that means that something else is blocking it from being abandoned. I'm pretty sure it still works in pruned mode as I'm pretty sure I did it before.

[-ck]

Abandon transaction is greyed out so presumably it's subject to the same limitations as zap.

No, that means that something else is blocking it from being abandoned. I'm pretty sure it still works in pruned mode as I'm pretty sure I did it before.

That's interesting in itself. What else could possibly be blocking it?

Code:

bitcoin-cli abandontransaction XX
error code: -5
error message:
Transaction not eligible for abandonment

[achow101]

That's interesting in itself. What else could possibly be blocking it?

Code:

bitcoin-cli abandontransaction XX
error code: -5
error message:
Transaction not eligible for abandonment

Anything that would cause this function: https://github.com/bitcoin/bitcoin/blob/02464da5e4aa8c19d4fff3859dcdee822e2af78c/src/wallet/wallet.cpp#L1052 to return false. So it seems the only criteria are that the transaction is not in the mempool and it is not confirmed. So clearing the mempool by restarting the node should work (if you are using any version greater than 0.13.2 i.e. a build of master, then you need to delete mempool.dat in the datadir).

[-ck]

That's interesting in itself. What else could possibly be blocking it?

Code:

bitcoin-cli abandontransaction XX
error code: -5
error message:
Transaction not eligible for abandonment

Anything that would cause this function: https://github.com/bitcoin/bitcoin/blob/02464da5e4aa8c19d4fff3859dcdee822e2af78c/src/wallet/wallet.cpp#L1052 to return false. So it seems the only criteria are that the transaction is not in the mempool and it is not confirmed. So clearing the mempool by restarting the node should work (if you are using any version greater than 0.13.2 i.e. a build of master, then you need to delete mempool.dat in the datadir).

Weird, I've restarted it a few times, perhaps another node that I have whitelisted elsewhere is submitting the transaction immediately once I restart. I'll try isolating it from the network and doing it.

[achow101]

Weird, I've restarted it a few times, perhaps another node that I have whitelisted elsewhere is submitting the transaction immediately once I restart. I'll try isolating it from the network and doing it.

If you run

Code:

getmempoolentry <txid>

beforehand and it tells you that the transaction is not in the mempool but you still can't abandon it, then there must be something else that is blocking it from being abandoned, but I don't know what that could be.

[-ck]

Weird, I've restarted it a few times, perhaps another node that I have whitelisted elsewhere is submitting the transaction immediately once I restart. I'll try isolating it from the network and doing it.

If you run

Code:

getmempoolentry <txid>

beforehand and it tells you that the transaction is not in the mempool but you still can't abandon it, then there must be something else that is blocking it from being abandoned, but I don't know what that could be.

Well it is in the mempool and it's 0.13.2 and restarting it doesn't remove it from the mempool... Hmm it could be because I getblocktemplates regularly from the bitcoind... or not :s

[achow101]

Well it is in the mempool and it's 0.13.2 and restarting it doesn't remove it from the mempool... Hmm it could be because I getblocktemplates regularly from the bitcoind... or not :s

Oh, duh, of course its in the mempool, restarting will have the wallet put the transactions back into the mempool on start. You need to set the -walletbroadcast=false command line option when you start it up again in order to prevent that from happening.

[-ck]

Well it is in the mempool and it's 0.13.2 and restarting it doesn't remove it from the mempool... Hmm it could be because I getblocktemplates regularly from the bitcoind... or not :s

Oh, duh, of course its in the mempool, restarting will have the wallet put the transactions back into the mempool on start. You need to set the -walletbroadcast=false command line option when you start it up again in order to prevent that from happening.

That did it. The abandon transaction option was no longer greyed out and now the transaction still appears in the interface but is marked in red and the details say not in mempool, even after restarting. I guess this is enough since theoretically it could still be bouncing around somewhere in the network at large. Thanks!

[-ck]

Hah the network had the last laugh and processed the transaction after about 3 days anyway.

[jnano]

The clunky interface is still around in the latest version.

"Abandon transaction" should never be grayed out. When it can't work, it should explain why. Or better, automatically stop broadcasting the wallet.

[achow101]

The clunky interface is still around in the latest version.

"Abandon transaction" should never be grayed out. When it can't work, it should explain why. Or better, automatically stop broadcasting the wallet.

Complaining about it here isn't going to help. Pretty much none of the Bitcoin Core developers read this forum (except for me). If you want to request a feature or change, open an issue in the github repo

[jnano]

I guess I'm just venting with no specific goal in mind.

Core's focus seems to be the network and RPC rather than the GUI. There are more important GUI features to deal with, so no point in bringing up this specific one.