Bitcoin Core v0.21.1 issue with incoming transaction

[wasisdn]

I have a issue loading my wallet.dat and found that the problem is with Core v0.21.0

Issue started yesterday when I opened Bitcoin Core, after receiving last payment from nicehash.
Always while loading the same block, I think the one in which the Nicehash transfer is located, Bitcoin Core was closed without an error message.
After trying several times to reinitialize, even from an older backup of my wallet.dat, I reinstalled v0.20.1 and the problem was resolved.

Bitcoin Client Software and Version Number: Bitcoin Core 0.21.0
Operating System: Windows 10 64-bit
System Hardware Specs: CPU Typ   HexaCore Intel Xeon X5690, 3600 MHz with 24GB RAM
Description of Problem: Bitcoin Core shut down while synching
Any Related Addresses: None
Any Related Transaction IDs: dad0f505c7c9697e1b7b79ecc7133701fd0a4424993ec27770191bfb6a290763
Screenshot of the problem: None
Log Files from the Bitcoin Client: https://pastebin.com/ZJk4tW0x

In the debug report shows 2 section
-v0.21.0 always the same issue and report
-v0.20.1 after reinstall without issue

[1715458947]

1715458947

[1715458947]

1715458947

[1715458947]

1715458947

[NotATether]

In the middle of the debug.log I see lines like:

Code:

2021-01-28T02:23:37Z UpdateTip: new best=00000000000000000006ddb5e854505f8b792122b0ac9469c07eb26db414f6fb height=667897 version=0x20006000 log2_work=92.622774 tx=610212544 date='2021-01-27T13:04:14Z' progress=0.999724 cache=439.9MiB(3354911txo)

Followed by bitcoin core initializing itself again:

Code:

2021-01-28T02:24:41Z Bitcoin Core version v0.20.1 (release build)
2021-01-28T02:24:41Z Qt 5.9.8 (static), plugin=windows (static)
2021-01-28T02:24:41Z System: Windows 10 (10.0), x86_64-little_endian-llp64

It means Bitcoin Core did not shut down cleanly either. Try to run 0.21.0 again but using -reindex and -rescan together. -rescan will fix your block.dat files in case they are corrupted, while -rescan will update your wallet file with newer transactions in the blockchain that are relevant to it.

You mentioned that this transaction ID is related:

6ce056a8-ae7c-44f3-9bf6-da7ac60c70e2

It's not showing up on block explorers when the dashes are removed.

Does 0.21.0 work if you use a different wallet file (or no wallet file at all)?

[wasisdn]

In the middle of the debug.log I see lines like:

Code:

2021-01-28T02:23:37Z UpdateTip: new best=00000000000000000006ddb5e854505f8b792122b0ac9469c07eb26db414f6fb height=667897 version=0x20006000 log2_work=92.622774 tx=610212544 date='2021-01-27T13:04:14Z' progress=0.999724 cache=439.9MiB(3354911txo)

Followed by bitcoin core initializing itself again:

Code:

2021-01-28T02:24:41Z Bitcoin Core version v0.20.1 (release build)
2021-01-28T02:24:41Z Qt 5.9.8 (static), plugin=windows (static)
2021-01-28T02:24:41Z System: Windows 10 (10.0), x86_64-little_endian-llp64

It means Bitcoin Core did not shut down cleanly either. Try to run 0.21.0 again but using -reindex and -rescan together. -rescan will fix your block.dat files in case they are corrupted, while -rescan will update your wallet file with newer transactions in the blockchain that are relevant to it.

You mentioned that this transaction ID is related:

6ce056a8-ae7c-44f3-9bf6-da7ac60c70e2

It's not showing up on block explorers when the dashes are removed.

Does 0.21.0 work if you use a different wallet file (or no wallet file at all)?

Sorry, above Transaction ID was given by Nicehash, maybe just an internal ID.

The correct ID is:

dad0f505c7c9697e1b7b79ecc7133701fd0a4424993ec27770191bfb6a290763

And yes, 0.21.0 it is working fine if I create a new wallet.dat.
It was working for a week or so untill this morning, the issue popup after reciving the first transaction last night to my wallet running Core version 0.21.0
I have also tried this way arround and import my wallet.dat after finished synching with a new wallet, but during inilizing my old wallet the same issue comes back, Bitcoin Core just shutdown without an error message and at same step in debug.log

[NotATether]

Please run bitcoin core 0.21.0 with the -debug switch and post the debug.log of that on a pastebin. There is not enough information inside to determine where the problem is.

[achow101]

There is nothing obviously wrong indicated in the debug.log file.

Can you post the contents of the db.log file? This can be found in the same directory as the wallet.dat file.

[wasisdn]

yes I know nothing much wrong, thats one of the reason why I open this thread.
The content of debug.log is shorten because of limited space, db.log is empty.
Actually I leave 2 part in the log;
Uper section when starting version 0.21.0 untill closed from alone while synching, always on same block without error message.
Lower section when starting bitcoin core older version 0.20.1 without problem.

Anyhow I reinstalled 0.21.0 again and start with -reindex and -rescan, now waiting for finish synching....

Thx a lot for your support, I will give feedback later

[darkv0rt3x]

Just to point out that Bitcoin Core can be compiled with native descriptor wallets (correct me if I'm wron) which are not compatible with legacy wallets.

So, maybe if you compiled it only with descriptor wallets, wouldn't it be supposed to not work well with old wallet format?

[wasisdn]

its definitly related to the last transaction from nicehash in combination with v0.21.0, the problem just start after that transaction.
Transaction ID:

dad0f505c7c9697e1b7b79ecc7133701fd0a4424993ec27770191bfb6a290763

v0.20.1 worked just fine for me again, but I reinstalled v0.21.0 to verify the problem.
Clean dbug.log and start with -reindex and -rescan.
After running for a couple of hours I went for dinner, synching sometime in 2017, when I came back Bitcon Core was shutdown.
Restart 2 times with same resault, the programm crashes without any error messege.

look at the debug.log, especially the entries in 800 and 1016 when programm crashes:

pastebin.com/mU1zjNDr

its the transaction from nicehash....

Now running v0.20.1 again, debug.log so far below, look at 206, its still synching without problem:

pastebin.com/wnA3cedH

[achow101]

Can you post the db.log file? Note that this is a different file from the debug.log file. It is named db.log and can be found in the same directory as the wallet.dat file.

[darkv0rt3x]

I can see this in his debug.log file:

Code:

2021-01-29T12:47:28Z ERROR: DeserializeFileDB: Failed to open file K:\Bitcoin qt\Database Wallet\anchors.dat

[achow101]

I can see this in his debug.log file:

Code:

2021-01-29T12:47:28Z ERROR: DeserializeFileDB: Failed to open file K:\Bitcoin qt\Database Wallet\anchors.dat

This is benign and expected.

What should happen if you stay at the version  v0.20.1  and not go to the v0.21.
I"m have started again about 3 weeks ago with the version v0.20.1 after troubles with my computer . Now i am still 4 years and 16 weeks behind. Block 235195 That goes very slow. I get the mension that it will still take 6 weeks to have completend the download.
I am affraid if i go to another version , i have to start completely again.

Please open a new thread for your issue. Do not hijack other people's threads.

[wasisdn]

Can you post the db.log file? Note that this is a different file from the debug.log file. It is named db.log and can be found in the same directory as the wallet.dat file.

yes I know there is a db.log but its empty

0.20.1 is now synchronized and runs without problem.
I will keep it for now as it is also faster connecting and syncing on startup, I just want to report this error.

[wasisdn]

still the same problem with version 0.21.1 whenever a new transaction comes in from Nicehash.
Bitcoin Core closes by itself without error message while synchronizing the latest transaction from Nicehash

Bitcoin Transaction: 6f8b84e3312abb77a95e5cb37c200cd564c6d4eb872b5d6a652fcb6eab70c5e9

debug.log

https://pastebin.com/eccGjEPT

start program new version 0.21.1
entry 169 programm closed without error message exactly with the latest transaction from nicehash

reinstall v0.20.1
entry 345 transaction from nicehash sync without problem...


any idea?

[nc50lc]

debug.log

2021-05-05T15:21:06Z Config file arg: dbcache="0"
.
.
.
2021-05-05T15:15:19Z * Using 0.5 MiB for block index database
2021-05-05T15:15:19Z * Using 1.8 MiB for chain state database
2021-05-05T15:15:19Z * Using 1.8 MiB for in-memory UTXO set (plus up to 286.1 MiB of unused mempool space)

-snip-

'0' is an invalid value for database cache, it might have been set to the minimum value of '4' instead.
Because of that, your node configured a very low in-memory UTXO set cache size.

Nicehash transactions are usually large in size because of batching and that may be the reason for the crash.
You can try setting dbcache to a high value, I'd suggest a quarter to half of your maximum RAM (depending on your typical usage).

[wasisdn]

debug.log

2021-05-05T15:21:06Z Config file arg: dbcache="0"
.
.
.
2021-05-05T15:15:19Z * Using 0.5 MiB for block index database
2021-05-05T15:15:19Z * Using 1.8 MiB for chain state database
2021-05-05T15:15:19Z * Using 1.8 MiB for in-memory UTXO set (plus up to 286.1 MiB of unused mempool space)

-snip-

'0' is an invalid value for database cache, it might have been set to the minimum value of '4' instead.
Because of that, your node configured a very low in-memory UTXO set cache size.

Nicehash transactions are usually large in size because of batching and that may be the reason for the crash.
You can try setting dbcache to a high value, I'd suggest a quarter to half of your maximum RAM (depending on your typical usage).

Thank you so much for the hint!
I was not aware of that and sounds very conclusive.
Now it looks like this:

2021-05-08T04:01:21Z Config file arg: dbcache="8192"
.
2021-05-08T04:01:21Z Cache configuration:
2021-05-08T04:01:21Z * Using 2.0 MiB for block index database
2021-05-08T04:01:21Z * Using 8.0 MiB for chain state database
2021-05-08T04:01:21Z * Using 8182.0 MiB for in-memory UTXO set (plus up to 286.1 MiB of unused mempool space)

I will post feedback here after receiving the next Nicehash transaction.
What I notice already, the synchronization is much faster with this setting...

[HCP]

What I notice already, the synchronization is much faster with this setting...

That's because Bitcoin Core is able to keep much more data in RAM instead of constantly read/writing it all out to disk...

You'll probably find the disk I/O is much reduced as well.

[wasisdn]

Unfortunately  same problem with the last Nicehash transaction today...

Transaction ID:

896521c7f5e216272905d82bdac301aa201a6733419f0f0e2b5c8e5b36eb25dc

debug.log

https://pastebin.com/EUaQj06N

v0.21.1
row 114 last entry before nicehash transaction and programm crashes

v020.1
row 240 same position as above but no crash
row 241 Nicehash transaction


Beside this problem with incomming transaction from nicehash v0.21.1 works without problem...

[nc50lc]

Maybe it's time to check for possible hardware issues since you log seems normal.
First check the disk drives, RAM and then the processor.

[wasisdn]

Maybe it's time to check for possible hardware issues since you log seems normal.
First check the disk drives, RAM and then the processor.

done already in the beginning, changed SSD, RAM and Power supply, only MB with CPU not.
But I can use this PC with Xeon CPU over days running 24/7, even CPU mining XMonero withhout any problem.
Funny that this problem has only occured since and with BitcoinCore v0.21.0 or above, all previous versions are ok, this PC running with BitcoinCore for more than 4 years now.
And also only with Nicehash incoming transactions, others are fine.

[wasisdn]

Since the debug.log doesn't show anything unusual & i notice that you're using Windows, you can try using "Event Viewer" software to find any error/warning around time when Bitcoin Core closed by itself.

If you never use it, check this guide https://www.howtogeek.com/school/using-windows-admin-tools-like-a-pro/lesson3/.

Thanks for your support, I will try my luck...


I really thought the latest BitcoinCore versions has a bug and someone might have the same problem or can tell what exactly is the difference from latest to previous versions so this problem could occur.
Can't imagine that it is due to my hardware or SW, since the problem is only with the incoming transactions from Nicehash and that only since Januar this year and since v0.21.0. This is for sure because if I re-syncronize the entire chain with latest BitcoinCore Version, it runs through all the years (many transaction over the past 4 years also incom from NH) until Januar 2021 without any problems and then always crashes on the same transaction from Nicehash dad0f505c7c9697e1b7b79ecc7133701fd0a4424993ec27770191bfb6a290763 (as in my first post). Then I revert to v0.20.1, sync this transaction (no problem), after that revert again to v0.21 and run until next transaction from Nicehash with same problem...
Maybe it has really to do with the size of transactions, because those from Nicehash are much larger than other, 7MB compared to usually a few hundred kb.
Now running with previous v0.20.1 and it works like a charm without any problems.

[nc50lc]

Maybe it has really to do with the size of transactions, because those from Nicehash are much larger than other, 7MB compared to usually a few hundred kb.

Most tx from Nicehash are just 7kB, about 6.8kvB in virtual size, 7MB wont even fit in a block.
But still, it's really weird that the problem only occurs when you receive transaction from them.

[wasisdn]

Maybe it has really to do with the size of transactions, because those from Nicehash are much larger than other, 7MB compared to usually a few hundred kb.

Most tx from Nicehash are just 7kB, about 6.8kvB in virtual size, 7MB wont even fit in a block.
But still, it's really weird that the problem only occurs when you receive transaction from them.

oh sorry yeah of course you are right 7kb vs few hundred byte.

Nicehash:
https://abload.de/img/snap2nljmn.jpg

other:
https://abload.de/img/snap319jxf.jpg

Could it related to my wallet address, because I use an older format address for Nicehash like 156... not bc1...?

[HCP]

Could it related to my wallet address, because I use an older format address for Nicehash like 156... not bc1...?

That shouldn't make any difference at all...


I am wondering, however, if this is related to this issue that @LoyceV found: https://bitcointalk.org/index.php?topic=5336919.0

Where Bitcoin Core is attempting to show all output addresses on the recent transactions list... and with your nicehash transaction having 200 odd outputs, it might be causing some sort of buffer overflow or "out of bounds" or similar error which is causing Bitcoin Core to shutdown?

It was noted in that LoyceV's thread, that because the list of addresses is truncated in the older versions of Bitcoin Core, that they work fine... which might also explain why you have noted that the old version is syncing/displaying those transactions without issue.

[nc50lc]

-snip-
I am wondering, however, if this is related to this issue that @LoyceV found: https://bitcointalk.org/index.php?topic=5336919.0

I think it's unrelated, that issue started in v0.21.1 while OP's started in v0.21.0.

On top of that, incoming transactions with multiple outputs that belong to your wallet will be displayed in separate entry per output and will only display that one address on 'recent transactions'.
Only those "send&receive" transactions (with in/out arrow icon, 'send to yourself') are one of the examples that still lists multiple addresses.

[wasisdn]

just to confirm after I ran BitcoinCore 0.21.1 without any problems for the entire last week until the next transaction from Nicehash arrived yesterday.
After the crash and several attempts with the same error at the same position (program closes when synchronizing the Nicehash transaction without a message) I pushed v0.20.1 into the program directory, started it and synchronized everything without problems.
Then switched back to the current v0.21.1 and its up and running again.

If you are interested, you can have a look at the last two days below, don't think I need to comment again

Transaction from NH is:

af4b27fd3e6917ddf39a16e159d2c20a4703ebaeffc554d3e29a780855a29525

Debug log:

https://pastebin.com/sWZwFZV0

I am currently running v0.21.1 and am waiting for a transaction from somewhere else.
Let's see if it goes smoothly...

[HCP]

Have you logged a description of this issue over on the Bitcoin Core github issue register? https://github.com/bitcoin/bitcoin/issues

It seems to me that you might get a better response from the developers by posting there.

There is nothing "obvious" in the debug as to why the app keeps shutting down when you receive a transaction from nicehash. Have you checked the Windows Event Viewer to see if there are any errors logged there for Bitcoin-Qt.exe?

Also, have you successfully received transactions from other sources? ie. Does Bitcoin Core shutdown when receiving any transaction... or does it only do it when you are receiving a transaction specifically from a nicehash payout?

[wasisdn]

Have you logged a description of this issue over on the Bitcoin Core github issue register? https://github.com/bitcoin/bitcoin/issues

It seems to me that you might get a better response from the developers by posting there.

There is nothing "obvious" in the debug as to why the app keeps shutting down when you receive a transaction from nicehash. Have you checked the Windows Event Viewer to see if there are any errors logged there for Bitcoin-Qt.exe?

Also, have you successfully received transactions from other sources? ie. Does Bitcoin Core shutdown when receiving any transaction... or does it only do it when you are receiving a transaction specifically from a nicehash payout?

Thanks for pointing me to the right place, just opened a new issue.
Nothing I can find in windows event related to BitcoinQT.
Still hesitating with TX from an exchange because I'm undecided about stocking up any more

[wasisdn]

Also, have you successfully received transactions from other sources? ie. Does Bitcoin Core shutdown when receiving any transaction... or does it only do it when you are receiving a transaction specifically from a nicehash payout?

Just transferred from an exchange and successfully synchronized with v0.21.1 ... so the problem is always and only with Nicehash transactions in connection with v0.21.x
 
txid

e2df3a45b07736c35eb391f883d47a7ec8e1c0cec2353b251984ecc90f99c024

row 109

http://https://pastebin.com/ysuUCwYp

[HCP]

Sorry, just to clarify... you're saying that line 109 in that debug output is the transaction arriving from an exchange... correct? And that your Bitcoin Core kept running just fine after receiving that transaction?

What's interesting is that going back and looking at your previous debug... the AddToWallet output is actually quite a long way from the debug that indicates that Bitcoin Core is shutting down, it seems to be syncing a lot of blocks after it sees that transaction.

Indeed the 2nd lot of debug here: https://pastebin.com/sWZwFZV0 near the very end, it doesn't even seem to be receiving any new wallet transaction at all, it is just syncing blocks and then shuts down (line 920 onwards)

[wasisdn]

Sorry, just to clarify... you're saying that line 109 in that debug output is the transaction arriving from an exchange... correct? And that your Bitcoin Core kept running just fine after receiving that transaction?

Yes exactly, v0.21.0 and tx from a source other than NH goes through without problems.

What's interesting is that going back and looking at your previous debug... the AddToWallet output is actually quite a long way from the debug that indicates that Bitcoin Core is shutting down, it seems to be syncing a lot of blocks after it sees that transaction.

That's because I didn't have BitcoinCore running for the week when I wasn't home, and when I knew that the tx from NH had been received, I started BitcoinCore with first attempt v0.21.1 and after crash with v0.20.1. Then of course the previous days will be processed first.

Indeed the 2nd lot of debug here: https://pastebin.com/sWZwFZV0 near the very end, it doesn't even seem to be receiving any new wallet transaction at all, it is just syncing blocks and then shuts down (line 920 onwards)

Correct, just to show that v0.21.0 runs without problems after I had previously syncronized the NH tx with v.0.20.1

line:
514 - start v0.21.1 attempt to sync NH tx
623 - crash just before NH tx 
2021-05-21T02:38:23Z UpdateTip: new best=000000000000000000076abd974283d6f1ac6ee46e6275289b919a0b72578fb9 height=684340 version=0x3fffe004 log2_work=92.890824 tx=643673329 date='2021-05-20T18:13:01Z' progress=0.999835 cache=22.4MiB(168883txo)

629 - restart with v0.20.1   
733 - same entry as above but no crash
2021-05-21T02:42:49Z UpdateTip: new best=000000000000000000076abd974283d6f1ac6ee46e6275289b919a0b72578fb9 height=684340 version=0x3fffe004 log2_work=92.890824 tx=643673329 date='2021-05-20T18:13:01Z' progress=0.999843 cache=22.4MiB(168885txo) warning='89 der letzten 100 Blöcke haben eine unerwartete Version'

734-735 - incoming tx from NH
806 - shut down

812-839 start - run - shutdown v0.21.1 without tx

The lines in second attempt are longer because a warning is appended to each line (unfortunately in German): warning = '89 of the last 100 blocks have an unexpected version'

[HCP]

That is indeed a puzzle... I am honestly all out of ideas for why this seems to be happening to you. I don't imagine you're the only Bitcoin Core user receiving payouts from Nicehash, so I'm surprised there are not more reports of this issue.

Hopefully the Bitcoin Core devs are able to investigate, identify the issue and fix it in a future version. For now, unless you absolutely require features in v0.21.1, I can only suggest running v0.20.1 until a new version is released and see if that fixes it.

[wasisdn]

That is exactly what surprises me too.
Anyway, I reported the issue on Github and someone already investigate. It seems that this is a new problem and nobody is using Bitcoincore for Win64 as an external wallet with Nicehash ... except me

Thank you very much for your support, I'll keep you informed!

[wasisdn]

Just to let you know, the latest version 22.0 has fixed the wallet problem, no more crashes on incoming transactions from anywhere.
Thanks again to everyone here in the forum and of course to the github guys!

[achow101]

Just to let you know, the latest version 22.0 has fixed the wallet problem, no more crashes on incoming transactions from anywhere.
Thanks again to everyone here in the forum and of course to the github guys!

That's kind of odd since the root cause was never pinned down and so no targeted attempt to fix the problem. But if it works, great!