Transactions are stuck in mempool if unconfirmed output is reused in another transaction.

[artemii235]

Can you reliably reproduce the issue?

If so, please list the steps to reproduce below:

1. Send the transaction.
2. Reuse it's unconfirmed output in another transaction.

Expected behaviour

All transactions mined.

Actual behaviour + errors

All transactions get stuck in some nodes mempool. E.g. these are visible in explorer, but not present on electrum servers, etc. The mempool is not synced between nodes.

The version of Komodo you were using:

Not sure about it, last error reported for OOT chain: https://discordapp.com/channels/412898016371015680/449950331577565194/661586034643042315
The transactions:
https://explorer.utrum.io/tx/d140693bc9c106ecfa28dd2d1cb3d5890371ee0dedd08f667f259faf9d8bf9b3
https://explorer.utrum.io/tx/bed8b7a4e6b42c1cdfdf689f8d279467eb266098dea7f585352b8b5349a3ea51
https://explorer.utrum.io/tx/90c9d04f6209825e2d42973e894d976cae0936514fd44cedd37fefac2359a449
were stuck unconfirmed until I explicitly rebroadcasted them using electrum server that didn't have them in mempool.

[jl777]

the best way is to just rebroadcasting the rawtx periodically until it confirms, ideally through a highly connected node.

for example, the caretaker network can run nodes that have the explorer and NN as peers. then when a new rawtx comes in that is not confirming after a few blocks, it can start sending the transaction to its local peers

[artemii235]

Do you mean MM2 caretakers concept by caretaker network or something else not related to MM2?

[jl777]

the concept of caretakers, meaning special nodes that some DEX holders will be running to keep the network running smoothly. they will also run my datablob layer to allow for easy way to broadcast whatever data that needs to get to all the caretaker nodes