wow ,almost 30,000 unconfirmed trans

[deisik]

increase the blocksize. duhhhhhhhhh.  idiots

I have yet to see it proved that it does actually have anything to do with the block size and not miners deliberately postponing confirmations, for whatever reason. Whoever is going to challenge that point, you know what to do. Basically, you should provide a list of the blocks mined since about time this congestion started with their respective sizes. The list is not going to be long, there are only 144 new blocks mined daily on average, and we will easily see who is right and who is not...

Anyone willing to compile such a list?

[Mr. Legendaris]

Unbelievable, why need 30.000 confirmations? I think this is not practice and make peoples leave bitcoin for transactions.

[Snorek]

Well, situation is getting better. When this thread was created yesterday we had almost 30.000 unconfirmed transactions pending.
When I checked out situation earlier today that number was more than 70.000 - now it is back to 30k level again.
It is downward trend or from now we will be stuck with plus 30k pending confirmation at all times?

[Meuh6879]

I call it a bitcoin fail, it just simply cannot scale to the current usage. Thanks.

I call it : ANTISPAM.
When you don't follow the (old) rules ... you hit the wall.

Networks not deal with this.
They apply rules.

[lol3c]

This is very unusual. I see that the number of transaction is still normal, but the number of unconfirmed transactions increases strangely. What might happen to the blockchain system ? Why do miners ignore their work ? Or is that because some problems appear on the network ?

[Meuh6879]

they don't ignore ... somes peoples pay more fees than other, it's all.




so, the others MUST WAIT.




it's nothing more than an antispam feature (rules MUST apply).

[erikalui]

Now I see my 2 days old transaction has been confirmed and earlier today when there were about 48k unconfirmed transactions, 10k transactions have been confirmed. The transactions with the recommended fee only are getting confirmed early compared to other transactions with a small fee.

[veleten]

this spam attack is good in a way that it will make some of the otherwise oblivious members educate themselves on how bitcoin tx fees work
guesstimating on a level of: "hey it is 500k so i pay 10k and if it is 0.05 I would splash 20k out" doesn't work and it is being punished by the prolonged wait
it does seem that an adequate (or slightly above) fee sees your tx confirmed no problem,just had a couple confirmed within 1 hour window
yes,it does take more money than usual to get high priority,but so is the current situation-deal with it
as much as it is frustrating to wait,please educate yourselves:

Predicting Bitcoin fees for transactions:

https://bitcoinfees.21.co/

http://bitcoin.stackexchange.com/questions/1195/how-to-calculate-transaction-size-before-sending

https://bitcointalk.org/index.php?topic=644189.0

[deisik]

this spam attack is good in a way that it will make some of the otherwise oblivious members educate themselves on how bitcoin tx fees work
guesstimating on a level of: "hey it is 500k so i pay 10k and if it is 0.05 I would splash 20k out" doesn't work and it is being punished by the prolonged wait

I'm afraid this might not have been a spam attack (whatever you might mean by this). Yesterday, there were quite a few blocks that had a lot free space in them (note, I didn't consider empty blocks), and this is in striking contrast with what miners are doing today and what they had been doing just before yesterday. Whoever might want to strongly disagree, may also want to check the blockchain for themselves...

So, it looks more like some miner attempting at getting more fees than someone spamming the network

[Carlton Banks]

yes,it does take more money than usual to get high priority,but so is the current situation-deal with it

Perception of "priority" is a real problem, thanks mostly to Blockchain.info.

Priority used to be based on how long ago a particular chunk of BTC was last spent: the older the chunk, the higher priority it has when it gets spent. This was to discourage spam, but also to encourage hoarders to spend.

But that logic has mostly been removed from Bitcoin, and it will be completely gone by the release of 0.14 (next major version being worked on now). But Blockchain.info haven't updated their "transaction" web page format to account for this, they're still using the previous age-based logic, and displaying it as if it's correct. Not very responsible for a big name in the industry, IMO.

[-ck]

I'm afraid this might not have been a spam attack (whatever you might mean by this). Yesterday, there were quite a few blocks that had a lot free space in them (note, I didn't consider empty blocks), and this is in striking contrast with what miners are doing today and what they had been doing just before yesterday. Whoever might want to strongly disagree, may also want to check the blockchain for themselves...

So, it looks more like some miner attempting at getting more fees than someone spamming the network

Well someone from the mining world has to clear this up so I'll explain.

What you are seeing from miners is a combination of clunky optimisations and bad configuration; it is not completely out of malice. Believe it or not, miners often do things without knowing they're doing them and are uninformed.  There are two major patterns to the blocks that are not full sized.

One is the "empty blocks" where there is only the generation transaction done as SPV (light wallet) mining without full validation of the block - these are done by many Chinese pools as a speed optimisation where they mine off the header of the previously found block temporarily from another pool until they can finish the validation completely and eventually switch to a fully validated block. They claim it is a way to save themselves from the delay inherent in the great firewall of China, though it is purely an optimisation to work around their less-scalable choice of pool software and setups. Most pools outside China do not use this. Some pooled mining software has been hacked to mine empty blocks even without use of header-only mining to speed up getting out block changes to its miners as well - there are other options though that are fast on block changes without this cludge but you can't force people to choose what software they run and how they run it.

See this for more discussion of empty blocks:
https://bitcointalk.org/index.php?topic=1085800.0

Second is purely a badly configured bitcoind either because they left it at the default or they configured it smaller by choice thinking it will provide them with a speed up to minimise their risk of orphans. The default in bitcoind is actually set to 750k and you can see many smaller "pools" or entities mining blocks of that size. Yes it's true, there are entities with millions of dollars worth of mining gear that don't know how to configure bitcoind. Additionally some pooled miners chose to actually set it lower simply to speed up work generation - notably p2pool users.

Yes the blocks would ALL be full if the miners configured their set ups correctly. The reality is these choices do speed up block generation and block propagation slightly and do decrease the risk of orphans, but with each next version of bitcoind these get smaller and smaller, and I maintain pooled mining software that is designed to be quick for fully validated and full sized block but of course I can't force everyone to use my software (even though it's free.)

Remember that fees are the long term incentive for miners to mine transactions into their block, but since the bulk of the fees on each block can be obtained by miners without actually filling the blocks, there is actually no major incentive for them to ensure each block is full unless they care about bitcoin transactions at large. Many miners DO care and choose their pools accordingly, but there are massive farms/entities that do not care or aren't even informed. They think they just need to get as much profit as quickly as possible from mining since the margins are now slim, and they are not aware of how their choices are affecting the network and bitcoin health. That doesn't mean their choices are actually leading to better profits by the way.

[deisik]

I'm afraid this might not have been a spam attack (whatever you might mean by this). Yesterday, there were quite a few blocks that had a lot free space in them (note, I didn't consider empty blocks), and this is in striking contrast with what miners are doing today and what they had been doing just before yesterday. Whoever might want to strongly disagree, may also want to check the blockchain for themselves...

So, it looks more like some miner attempting at getting more fees than someone spamming the network

Well someone from the mining world has to clear this up so I'll explain.

What you are seeing from miners is a combination of clunky optimisations and bad configuration; it is not completely out of malice. Believe it or not, miners often do things without knowing they're doing them and are uninformed.  There are two major patterns to the blocks that are not full sized

Don't worry, I read the rest of your post and just saving screen space

In this way, miners are deliberately (and some part of them maybe through ignorance) pursuing profits in the form of the mining reward only, not caring much about fees collected and transactions included, thus the empty blocks. So far so good. But this doesn't explain why some blocks are only half full (and I don't mean those 750k blocks). For example, this block has only 82 transactions. A few other blocks have a few hundred transactions in them, but they are still less than half full (this and this blocks). And such blocks were seen only yesterday, there were no such blocks today or before yesterday. What's the catch really?

[Meuh6879]

Looking good ... purge in progress.

[ccminer.net]

Looking good ... purge in progress.

where did you get this plot ?

[Meuh6879]

  read the logo in watermark.

[-ck]

I'm afraid this might not have been a spam attack (whatever you might mean by this). Yesterday, there were quite a few blocks that had a lot free space in them (note, I didn't consider empty blocks), and this is in striking contrast with what miners are doing today and what they had been doing just before yesterday. Whoever might want to strongly disagree, may also want to check the blockchain for themselves...

So, it looks more like some miner attempting at getting more fees than someone spamming the network

Well someone from the mining world has to clear this up so I'll explain.

What you are seeing from miners is a combination of clunky optimisations and bad configuration; it is not completely out of malice. Believe it or not, miners often do things without knowing they're doing them and are uninformed.  There are two major patterns to the blocks that are not full sized

Don't worry, I read the rest of your post and just saving screen space

In this way, miners are deliberately (and some part of them maybe through ignorance) pursuing profits in the form of the mining reward only, not caring much about fees collected and transactions included, thus the empty blocks. So far so good. But this doesn't explain why some blocks are only half full (and I don't mean those 750k blocks). For example, this block has only 82 transactions. A few other blocks have a few hundred transactions in them, but they are still less than half full (this and this blocks). And such blocks were seen only yesterday, there were no such blocks today or before yesterday. What's the catch really?

Pretty sure I explained - dodgy optimisation choices in a hacked coin daemon. It was notably less than 2 minutes after the previous block. The mempool would have clearly been full in a regular bitcoind at the time so there's no way a call to create a block template would have only given them a 33kB block.

EDIT: Look at the block sizes here for the last 30 days and sort by average size:
http://data.bitcoinity.org/bitcoin/blocksize/30d?t=l

You'll see that famously the smallest average block size is by Eligius which is the pool associated with Luke-jr... So that one's clearly not a conspiracy by the people pushing for bigger blocks. 3 of the top 4 average sizes are from ckpool users.

[veleten]

yes,it does take more money than usual to get high priority,but so is the current situation-deal with it

Perception of "priority" is a real problem, thanks mostly to Blockchain.info.

Priority used to be based on how long ago a particular chunk of BTC was last spent: the older the chunk, the higher priority it has when it gets spent. This was to discourage spam, but also to encourage hoarders to spend.

But that logic has mostly been removed from Bitcoin, and it will be completely gone by the release of 0.14 (next major version being worked on now). But Blockchain.info haven't updated their "transaction" web page format to account for this, they're still using the previous age-based logic, and displaying it as if it's correct. Not very responsible for a big name in the industry, IMO.

didn't know that,thank you for clarification
besides blockchain.info estimation of optimal fees is usually scewed and not in the senders favour,it tends to suggest ridiculously high fees

[Carlton Banks]

besides blockchain.info estimation of optimal fees is usually scewed and not in the senders favour,it tends to suggest ridiculously high fees

Yes, that's the impression I get from other BC.i wallet users too. If they can tell you (accurately) what the size of your transaction will be before you set the fee, you can use https://bitcoinfees.21.co (and a little old-fashioned arithmetic) to work out what a sensible fee should be.

[countryfree]

The miners are doing everything they can, but that's clearly not enough.
Check this one:

https://btc.com/000000000000000002e4403e2babdccfe89704277ea48e1ab3ed5957cdfb1f24

That block contains 2,469 transactions (is this the new record?), and it's nearly touching the 1MB block size limit at 998,085 Bytes.

I'm still waiting for larger block sizes, or SegWit which could assuredly help on the short term, but meanwhile there is no other solution than to go for larger fees, with the risk of making BTC less popular.

Miners, and all BTC developers should understand that BTC's growth requires a growing infrastructure and includes a growing block size. Visa wasn't able to handle thousands of transactions every second some decades ago, but Visa has good managers and they anticipated the growth of the network. BTC needs to do the same. If I'm reasonable, I'd say the block size should be increased 50% each year.

Nope, saw one by HaoBTC for 3200 transactions yesterday,

Theoretical limit is 4200 transactions per block.

 

3200 transactions, that's impressive, and it's also a matter of chance, as there were very few large transactions in that block.
Things look like it's getting better with only 27K transactions waiting as of now, but there might plenty of large ones in that number...

Made a SEPA transfer from Switzerland to the UK last Friday morning. The money arrived on Tuesday morning. That's awfully slow but that transfer was free.

[Carlton Banks]

I'm still waiting for larger block sizes, or SegWit which could assuredly help on the short term

Segwit is an increase in the Blocksize. There is no "larger blocks, or Segwit"

Miners, and all BTC developers should understand that BTC's growth requires a growing infrastructure and includes a growing block size. Visa wasn't able to handle thousands of transactions every second some decades ago, but Visa has good managers and they anticipated the growth of the network. BTC needs to do the same. If I'm reasonable, I'd say the block size should be increased 50% each year.

What makes you think that blocksize changes are the only way to increase the transaction rate? If you were reasonable, you'd say that blocksize increases should be a last resort, and that anything else that improves the transaction rate should be implemented before that.