Author

Topic: Unconfirmed transactions (Read 2314 times)

member
Activity: 77
Merit: 10
April 29, 2013, 08:35:16 PM
#19
Here's a recent example:
A guy's been collecting spam at 13y4nywLxwAVphCrocjt69677cvMjZ28cM and is now trying to spend it.
It was a 45k transaction..  my bitcoind node shot it down :
2013-04-28 19:21:09 ERROR: CTxMemPool::accept() : not enough fees 6796fb03ec659cab640a3bd78f9a50901604cf18de08ca51b38c09a38a721f86, 440000 < 460000
but it's made it as far as blockchain.info
https://blockchain.info/tx/6796fb03ec659cab640a3bd78f9a50901604cf18de08ca51b38c09a38a721f86

According to blockchain.info, he's converted 0.00529172btc into 0.00089172 btc (0.0044 fees) and even then there wasn't enough fees to satisfy the p2p network.

That's not a good user experience.

Looks like all this faucet spam is in limbo still.  Whatever wallet they used miscalculated the fees.  It looks like they were aiming for 0.0001/kb and rounded down instead of up.  They needed to send 0.0046 btc fees instead of 0.0044 btc and then btcguild would likely have mined it.

You can see all the input spam with advanced mode.  https://blockchain.info/tx/6796fb03ec659cab640a3bd78f9a50901604cf18de08ca51b38c09a38a721f86?show_adv=true
hero member
Activity: 784
Merit: 1000
0xFB0D8D1534241423
April 28, 2013, 06:50:43 PM
#18
I've always paid a fee of 0.01BTC regardless of what the settings were. So 0.01*133(Current price) = $1.33
That's a huge fee. I sent this 50+ input tx for free recently.
A 0.01 BTC fee is only necessary for a 200KB tx... we're talking more than a thousand inputs.
Let me reiterate
200 freaking kilobytes.
1300+ freaking inputs.

For amounts less than $40 or so, you're doing worse than the merchant would with PayPal.
Pay freaking Pal.
A lot of those unconfirmed transaction are micro-faucet spam.  If the block size was increased they might go through easier.
148 "inputs" = 148000 bytes. (nope, is larger than 10k)
You're off by more than half an order of magnitude, not that it changes the result.

Indeed... 148 inputs vs 1000 inputs.. gah.  Still.. 148*1000 == 1000*148  as you pointed out.

In any case, the source I used was: http://bitcoin.stackexchange.com/questions/1195/how-to-calculate-transaction-size-before-sending/3011#3011
I know it's a generalization but if it's way off the mark I'd love to know.
It's close enough. It's spot on. To provide a nice cushion, I use 250 bytes/input. (Really, it's about the mental math. One bitcoin-day per input. Simple.)
member
Activity: 77
Merit: 10
April 28, 2013, 06:11:47 PM
#17
A lot of those unconfirmed transaction are micro-faucet spam.  If the block size was increased they might go through easier.
148 "inputs" = 148000 bytes. (nope, is larger than 10k)
You're off by more than half an order of magnitude, not that it changes the result.

Indeed... 148 inputs vs 1000 inputs.. gah.  Still.. 148*1000 == 1000*148  as you pointed out.

In any case, the source I used was: http://bitcoin.stackexchange.com/questions/1195/how-to-calculate-transaction-size-before-sending/3011#3011
I know it's a generalization but if it's way off the mark I'd love to know.
legendary
Activity: 1862
Merit: 1011
Reverse engineer from time to time
April 28, 2013, 06:09:46 PM
#16
I've always paid a fee of 0.01BTC regardless of what the settings were. So 0.01*133(Current price) = $1.33
hero member
Activity: 784
Merit: 1000
0xFB0D8D1534241423
April 28, 2013, 05:54:09 PM
#15
I think that these unconfirmed transactions is going to be a big key factor that keeps more people from adopting this as a more widespread currency. Yes while it's true if you include a fee it is more likely to get accepted quickly but it seems to me that non-technical people trying to adopt to bitcoin will be frustrated by this...having transactions unconfirmed for hours or days, or transactions that are lost forever. Thoughts on this?

are you telling me that a 0.01$ fee for transaction is too much?

Or even less...?
If you do the math, it's actually waay over a dollar per transaction.
Again with the orders of magnitude...
$150 (bullish, for convenience of math) * 0.0005 = $0.075
Alternatively,
$1.00 / $150 = BTC0.06666667
That's a huge TX fee, and it corresponds to a 13KB transaction, generally only attainable with faucet / SatoshiDICE spam.
legendary
Activity: 1862
Merit: 1011
Reverse engineer from time to time
April 28, 2013, 05:44:26 PM
#14
I think that these unconfirmed transactions is going to be a big key factor that keeps more people from adopting this as a more widespread currency. Yes while it's true if you include a fee it is more likely to get accepted quickly but it seems to me that non-technical people trying to adopt to bitcoin will be frustrated by this...having transactions unconfirmed for hours or days, or transactions that are lost forever. Thoughts on this?

are you telling me that a 0.01$ fee for transaction is too much?

Or even less...?
If you do the math, it's actually waay over a dollar per transaction.
hero member
Activity: 784
Merit: 1000
0xFB0D8D1534241423
April 28, 2013, 05:38:52 PM
#13
A lot of those unconfirmed transaction are micro-faucet spam.  If the block size was increased they might go through easier.
148 "inputs" = 148000 bytes. (nope, is larger than 10k)
You're off by more than half an order of magnitude, not that it changes the result.
member
Activity: 77
Merit: 10
April 28, 2013, 05:28:15 PM
#12
A lot of those unconfirmed transaction are micro-faucet spam.  If the block size was increased they might go through easier.

Nope.  Unspendable faucet spam is one of the bigger recent tragedies.   Where "unspendable" = you have to get more bitcoin from somewhere else to be able to spend what you got.  People think of addresses as "accounts" and that multiple small transactions add up, which they do not.

To spend something for free, the output has to be >= 0.01, and the transaction has to be less than 10k.  What is one to do with a 0.00001 btc like you'd get from bitcoinforrest?  Suppose we let them stack up:

Supposing that each input costs roughly 148 bytes to spend (after compressed public keys) and suppose somebody collected 1000 x 0.00001 faucet spams then we have roughly:

1000 x 0.00001 = 0.01 btc. (yep, meets >= 0.01 minimum)
1000 "inputs" x 148 = 148000 bytes. (nope, is larger than 10k)
Fees according to https://en.bitcoin.it/wiki/Transaction_fees are required, its more than 10kb:  148 x 0.0005 = 0.074 btc fees.
The p2p network will relay the transactions as low as 0.0001 btc per kb, so the bare minimum is 148 x 0.0001 = 0.0148 btc fees.

ie: if you got no btc from any other sources, you can't spend the spam.

It's not a blocksize issue at all.  It is this:
2013-04-25 18:27:48 ERROR: CTxMemPool::accept() : not enough fees 9b121b407c6722a331b46e71cb0cc922ac8eb8ef1ee38f99ce6c816e3483f6b3, 0 < 10000
The peer-to-peer network is refusing to relay the transaction.  There are good odds it won't ever reach a miner, even if they would mine it for free.
 
And yet we get new "services" like bitcoinforrest exploiting new users.  It leads to frustration and people just giving up and deleting wallets.

Not to mention that spending this spam would consume half a typical miner block..

Here's a recent example:
A guy's been collecting spam at 13y4nywLxwAVphCrocjt69677cvMjZ28cM and is now trying to spend it.
It was a 45k transaction..  my bitcoind node shot it down :
2013-04-28 19:21:09 ERROR: CTxMemPool::accept() : not enough fees 6796fb03ec659cab640a3bd78f9a50901604cf18de08ca51b38c09a38a721f86, 440000 < 460000
but it's made it as far as blockchain.info
https://blockchain.info/tx/6796fb03ec659cab640a3bd78f9a50901604cf18de08ca51b38c09a38a721f86

According to blockchain.info, he's converted 0.00529172btc into 0.00089172 btc (0.0044 fees) and even then there wasn't enough fees to satisfy the p2p network.

That's not a good user experience.
cp1
hero member
Activity: 616
Merit: 500
Stop using branwallets
April 28, 2013, 02:09:52 PM
#11
A lot of those unconfirmed transaction are micro-faucet spam.  If the block size was increased they might go through easier.
hero member
Activity: 616
Merit: 500
April 28, 2013, 12:45:06 PM
#10
so how long will a transaction just sit unconfirmed before the sender gets access to their funds back?

i've had 2 btc in limbo since about 24 hours ago.

That depends on which wallet you used to send it.

Bitcoin-Qt will continue to try to relay the transaction forever.  While it is unlikely, it is possible that a transaction sent via Bitcoin-Qt with no fee will continue to sit unconfirmed after several years.

https://blockchain.info/wallet has a expiration, I'm not sure, but I think it's about 48 hours.  If you send a transaction with the blockchain.info wallet, and it is still unconfirmed after the expiration, blockchain.info will stop trying to relay the transaction and will return the funds to your wallet.

I'm not sure how Electrum or MultiBit handle the situation.


it was blockchain. thanks for the answer.

legendary
Activity: 3472
Merit: 4801
April 28, 2013, 11:28:43 AM
#9
so how long will a transaction just sit unconfirmed before the sender gets access to their funds back?

i've had 2 btc in limbo since about 24 hours ago.

That depends on which wallet you used to send it.

Bitcoin-Qt will continue to try to relay the transaction forever.  While it is unlikely, it is possible that a transaction sent via Bitcoin-Qt with no fee will continue to sit unconfirmed after several years.

https://blockchain.info/wallet has a expiration, I'm not sure, but I think it's about 48 hours.  If you send a transaction with the blockchain.info wallet, and it is still unconfirmed after the expiration, blockchain.info will stop trying to relay the transaction and will return the funds to your wallet.

I'm not sure how Electrum or MultiBit handle the situation.
hero member
Activity: 616
Merit: 500
April 28, 2013, 11:18:56 AM
#8
so how long will a transaction just sit unconfirmed before the sender gets access to their funds back?

i've had 2 btc in limbo since about 24 hours ago.
legendary
Activity: 3472
Merit: 4801
April 08, 2013, 06:56:23 PM
#7
Add 1 shatoshi and you will have  6 confirmation fast.

Perhaps a bit more than 1 satoshi.  Most miners will treat 1 satoshi the same as 0.

Consider adding a 0.0005 BTC fee.  (That's about $0.09 at today's exchange rate).
full member
Activity: 161
Merit: 100
April 08, 2013, 11:32:33 AM
#6
Add 1 shatoshi and you will have  6 confirmation fast.
full member
Activity: 167
Merit: 100
April 08, 2013, 10:47:25 AM
#5
I currently have 4 transactions incoming that haven't been confirmed yet... 3 from April 6th in the afternoon and one from April 7th in the morning. Nothing big... just some faucet transactions. Still seems strange that they are taking so long to confirm. Another thread mentioned that there were close to 12,000 unconfirmed transactions about 12 hours ago, but checking the Blockchain unconfirmed transactions page now shows only about 350-400.
member
Activity: 112
Merit: 10
April 03, 2013, 06:09:32 PM
#4
I have my blockchain wallet set to include the maximum fee on each transaction, but two transactions I sent out this morning were stripped of their fees and remain unconfirmed. 
member
Activity: 106
Merit: 10
April 03, 2013, 06:05:43 PM
#3
Are people really smart enough to know what type of fee they need to include? I actually haven't sent very many payments myself. Is the main client's "automatic fee" good enough to get the transaction confirmed in a reasonable amount of time? What about larger-value transactions that need no fee? People generally aren't going to include an optional fee unless they have to, then they'll get stuck waiting.
hero member
Activity: 490
Merit: 500
April 03, 2013, 03:51:54 PM
#2
I think that these unconfirmed transactions is going to be a big key factor that keeps more people from adopting this as a more widespread currency. Yes while it's true if you include a fee it is more likely to get accepted quickly but it seems to me that non-technical people trying to adopt to bitcoin will be frustrated by this...having transactions unconfirmed for hours or days, or transactions that are lost forever. Thoughts on this?

are you telling me that a 0.01$ fee for transaction is too much?

Or even less...?
member
Activity: 106
Merit: 10
April 03, 2013, 03:07:03 PM
#1
I think that these unconfirmed transactions is going to be a big key factor that keeps more people from adopting this as a more widespread currency. Yes while it's true if you include a fee it is more likely to get accepted quickly but it seems to me that non-technical people trying to adopt to bitcoin will be frustrated by this...having transactions unconfirmed for hours or days, or transactions that are lost forever. Thoughts on this?
Jump to: