Bitcoind crashed and no longer starts

[tuaris]

At exactly 8:30PM EST my bitcoind service crashed (out of nowhere) after running for several months:

Code:

Dec 24 20:30:39 bitcoin kernel: pid 3664 (bitcoind), uid 0: exited on signal 6 (core dumped)

Now it will no longer start, it dies with this message:

Code:

Assertion failed: (internal_key.size() >= 8), function ExtractUserKey, file ./db                                           /dbformat.h, line 96

The debug log only shows:

Code:

2013-12-25 01:49:00 Bitcoin version v0.8.5.0-gef14a26-beta ()
2013-12-25 01:49:00 Using OpenSSL version OpenSSL 0.9.8q 2 Dec 2010
2013-12-25 01:49:00 Default data directory /.bitcoin
2013-12-25 01:49:00 Using data directory /var/db/bitcoin/
2013-12-25 01:49:00 Using at most 125 connections (11095 file descriptors available)
2013-12-25 01:49:00 Using 2 threads for script verification
2013-12-25 01:49:00 init message: Verifying wallet...
2013-12-25 01:49:00 dbenv.open LogDir=/var/db/bitcoin/database ErrorFile=/var/db/bitcoin/db.log
2013-12-25 01:49:00 Bound to [::]:8333
2013-12-25 01:49:00 Bound to 0.0.0.0:8333
2013-12-25 01:49:00 init message: Loading block index...
2013-12-25 01:49:00 Opening LevelDB in /var/db/bitcoin/blocks/index
2013-12-25 01:49:00 Opened LevelDB successfully
2013-12-25 01:49:00 Opening LevelDB in /var/db/bitcoin/chainstate
2013-12-25 01:49:00 Opened LevelDB successfully

This is bitcoin 0.8.5 on FreeBSD 9.0 32-bit.  Disk space and RAM are okay.

[1714935066]

1714935066

[1714935066]

1714935066

[1714935066]

1714935066

[tuaris]

Upgrading to 0.8.6 does not fix this problem.

[Doge Road]

Show db.log.

[tiktoc]

Wallet might be corrupted at a guess from the error output. If it DOES NOT have coins in it then delete the wallet and try again.

[DustyRah]

All of my clients don't work after upgrading to Windows 8.1 from Windows 8.

They start up and immediately give error saying "bitcoin-qt.exe has stopped working" and similarly "namecoin-qt.exe has stopped working" etc.

I have tried totally uninstalling and installing again after removing from both install folder and Apps\Roaming folders. Also tried to start with -reindex parameter but no luck. Tried as administrator as well.

Not sure what's up but all qt based apps don't work now on Windows 8.1.

Anyone have any suggestions please help me out!!

Thanks for your help~

[U1TRA_L0RD]

First back up anything bitcoin related thats important, and then reinstall windows and reinstall those files.

[evilpete]

At exactly 8:30PM EST my bitcoind service crashed (out of nowhere) after running for several months:

Code:

Dec 24 20:30:39 bitcoin kernel: pid 3664 (bitcoind), uid 0: exited on signal 6 (core dumped)

Now it will no longer start, it dies with this message:

Code:

Assertion failed: (internal_key.size() >= 8), function ExtractUserKey, file ./db                                           /dbformat.h, line 96

The debug log only shows:

Code:

2013-12-25 01:49:00 Bitcoin version v0.8.5.0-gef14a26-beta ()
2013-12-25 01:49:00 Using OpenSSL version OpenSSL 0.9.8q 2 Dec 2010
2013-12-25 01:49:00 Default data directory /.bitcoin
2013-12-25 01:49:00 Using data directory /var/db/bitcoin/
2013-12-25 01:49:00 Using at most 125 connections (11095 file descriptors available)
2013-12-25 01:49:00 Using 2 threads for script verification
2013-12-25 01:49:00 init message: Verifying wallet...
2013-12-25 01:49:00 dbenv.open LogDir=/var/db/bitcoin/database ErrorFile=/var/db/bitcoin/db.log
2013-12-25 01:49:00 Bound to [::]:8333
2013-12-25 01:49:00 Bound to 0.0.0.0:8333
2013-12-25 01:49:00 init message: Loading block index...
2013-12-25 01:49:00 Opening LevelDB in /var/db/bitcoin/blocks/index
2013-12-25 01:49:00 Opened LevelDB successfully
2013-12-25 01:49:00 Opening LevelDB in /var/db/bitcoin/chainstate
2013-12-25 01:49:00 Opened LevelDB successfully

This is bitcoin 0.8.5 on FreeBSD 9.0 32-bit.  Disk space and RAM are okay.

Exactly the same here (except openssl 1.0.1e and 64 bit 11.0).  I tried a reindex but found I got the same errors during reindex that the Mac OSX folks typically had.  eg:  Invalid argument: not an sstable (bad magic number)

I looked at the 0.8.5 -> 0.8.6 patch to env_posix.cc that disabled mmap() for writing in LevelDB for OSX.  When I enabled that for FreeBSD as well, the reindex problem was solved.