I have three transactions that are linked, like the output of the first one feeds into the second, and the second into the third. RBF is on. All three are still unconfirmed in the mempool. I want to swap them out for a new set of three, each with a fee that’s one satoshi more than the originals. Can I do that, or does the first transaction in the new set have to outbid all three of the old ones?
Can RBF Replace Multiple Transactions in the Mempool?
10 replies 436 views
chrisomegaFull Member
Posts: 69 · Reputation: 631
#2Mar 15, 2017, 05:38 PM
You can totally replace any of those transactions. If you swap out the first one, the second and third will get kicked out too. Change the second one, and the first stays but the third goes. Replace the third, and the first two stick around.
I see what you’re saying, but that’s tricky for me. I want to replace all three. For instance, my first set is like this: TX11 (9 sat/B) -> TX21 (9 sat/B) -> TX31 (9 sat/B). Now I want to change it to TX12 (10 sat/B) -> TX22 (10 sat/B) -> TX32 (10 sat/B). Wouldn't the miner want to accept my new one since it pays more?
Your graph is a bit confusing. Where’s TX12 to TX20 and TX22 to TX30? If you mean the right one is a child of the left, it’s clearer to phrase it like this: TX1A (9 sat/B) -> TX2A (9 sat/B) -> TX3A (9 sat/B). Switch A to B for the new one. But it’s not just 30 sat/vB. You gotta cover the total fees from the children too.
wizard_2016Full Member
Posts: 46 · Reputation: 575
#5Mar 18, 2017, 02:59 AM
What about child-pays-for-parent? Like, RBF the last transaction with enough fee to cover its unconfirmed parents.
chrisomegaFull Member
Posts: 69 · Reputation: 631
#6Mar 18, 2017, 07:13 AM
Yeah, I recommended that too. But OP seems set on replacing all three individually, maybe with new ones going to different addresses.
No way. This isn’t just about payments. Those graphs are for a layer 2 thing. Each transaction has an OP_Return output and a change address for making a new token. But if miners don’t include them in a block, they kinda expire for layer 2. So, they need RBF.
Why do it on Bitcoin then? If it’s just data pushing, testnets are way better. Or make your own chain and finalize stuff cheaply.
Bitcoin is the stronghold. Plus my layer 2 works on voting with tokens. More tokens mean more power, kinda like PoW. If they don’t get into a Bitcoin block, I lose the fees I paid. So yeah, RBF helps.
I thought RBF would work like this: I send an "inv" message for my three transactions. When a node gets it, they know all three come together. So when they ask for the data, I’d send them back in three separate "tx" messages. No bulk option, you know?
I figured out I can change my layer 2 algorithm to ditch the need for RBF. Thanks for all the insights, though!
Related topics
- Clipboard Vulnerabilities in Cryptocurrency Transactions 8
- Exploring AI Applications in Bitcoin's Mempool 0
- New Bitcoin Improvement Proposal with $100 Reward 9
- Understanding the Differences Between Traditional and Simplified Chinese Mnemonics 6
- Understanding Fees with Taproot Script Usage 3
- Can You Prune Bitcoin Core Data by Date Range? 4