So I was finally able to get Dogecoin working. I was using the wrong magic. I have everything up on a live server now and it doesn't seem to be polling the database for the queries, instead it is opening up the block file every single time.
Please search for "no-serve" in this thread. |
|
|
|
What would be faster, using one DB for all the blocks
or each chain one DB
I guess you should use a separate database per chain. It complicates the HTTP side but simplifies deleting a corrupt (or abandoned) chain and spreading an increasing load. I don't know if it would be much faster. |
|
|
|
block_tx 1181 1
Skipped 95 bytes at block end
Exception at 235523
Failed to catch up {'blkfile_number': 100000, 'dirname': u'/home/ubuntu/.dogecoin', 'chain_id': 8, 'id': 1, 'blkfile_offset': 208572}
Traceback (most recent call last):
File "/usr/local/lib/python2.7/dist-packages/Abe/DataStore.py", line 2422, in catch_up
store.catch_up_dir(dircfg)
File "/usr/local/lib/python2.7/dist-packages/Abe/DataStore.py", line 2475, in catch_up_dir
store.import_blkdat(dircfg, ds, filename[0])
File "/usr/local/lib/python2.7/dist-packages/Abe/DataStore.py", line 2596, in import_blkdat
b = store.parse_block(ds, chain_id, magic, length)
File "/usr/local/lib/python2.7/dist-packages/Abe/DataStore.py", line 2630, in parse_block
d['transactions'].append(deserialize.parse_Transaction(ds))
File "/usr/local/lib/python2.7/dist-packages/Abe/deserialize.py", line 90, in parse_Transaction
for i in xrange(n_vout):
OverflowError: Python int too large to convert to C long Most likely the transaction format differs from Bitcoin's. NovaCoin and CryptoCash also changed their transaction format, and Abe/Chain.py accommodates them (by passing has_nTime in PpcPosChain.ds_parse_transaction). If Doge shared their format, it could do the same, but it seems different, since the default parser got to Block 1181. It is a matter of translating the C++ definition of CTransaction into Python. |
|
|
|
Thank you for the quick update! I pulled the fix and reran Abe but the same error still appears.
If the error is still at File "/usr/local/lib/python2.7/dist-packages/Abe/deserialize.py", line 247, in script_GetOp then Python is using the old, installed version. Either reinstall Abe or adjust PYTHONPATH to use the new code. |
|
|
|
When no-statistics Abe starts working on Bitcoin blocks in late Oct 2013, it is showing a IndexError: string index out of range. Retrying does not help the problem. Any ideas how to solve this?
I've just copied a fix from the master branch that I think handles this. Please pull and retry. Thanks. |
|
|
|
Hello, Sorry for not Being Precise, Here is the Output in the Command Line when I visit the Block Chain Explorer Website: Traceback (most recent call last):
File "/usr/lib/python2.7/wsgiref/handlers.py", line 85, in run
self.result = application(self.environ, self.start_response)
File "/root/bce/bitcoin-abe/Abe/abe.py", line 209, in __call__
handler(page)
File "/root/bce/bitcoin-abe/Abe/abe.py", line 307, in handle_chains
100.0 - (100.0 * (ss + more) / (total_ss + more)))
ZeroDivisionError: float division by zero
183.87.156.69 - - [28/Feb/2014 12:25:58] "GET / HTTP/1.1" 500 59 This looks like an older version of bitcoin-abe. Can you please try the latest from https://github.com/bitcoin-abe/bitcoin-abe/ ? |
|
|
|
Please verify the line that you added to /etc/sudoers.
No longer getting tty, still getting the fastcgi errors... [Fri Feb 28 11:53:21 2014] [warn] [client 108.162.237.97] (104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server
[Fri Feb 28 11:53:21 2014] [error] [client 108.162.237.97] Premature end of script headers: abe.fcgi
Trying to use FastCGI as outlined...having issues. Sorry, but I can not troubleshoot the FastCGI setup here. Does Abe work in non-FCGI mode with --port=2750? Are your scripts executable? Can you find the output of abe.fcgi? It may include a useful error message. |
|
|
|
Somehow the current block height displayed on Abe's main page no longer changes as new blocks are processed by Abe. Right now the current block displayed on main page is 521941. However I am able to access a newer block 521948 by going to http://localhost:2750/block/a4436cbcf8986ab38e64fc88983c973a76cd938157dbb2f0d3a20748842dc8d5. Abe's title also shows Block a443...48842dc8d5 instead of Block 521948 q/getblockcount also shows the wrong block height of 521941. Is it possible to do like a rescan to update the Abe's current block height on the front page? The chain was probably forked. Once the current chain catches up to the previous highest block it will sort itself out. I've seen this a few times. Abe keeps orphan blocks and side chains and reuses them if they join the main chain. Thanks for the quick reply! The chain I'm using is LTC, is this just a temp fork that affects some people, and also the reason why we wait for several confirmations for a transaction? Well, this looks to me like Abe failing to find block 521941 by its hash in block 521942. This can occur because of bugs, especially when two or more processes simultaneously load blocks. This can also occur because of block format changes, such as changes to the hash algorithm. Can you browse 521942 by its hash? You might have to delete 521941 and/or 521942 from the database and rescan the blockfile that contains them. I don't have a tool to do exactly that, but you may be able to avoid a full reload with this, assuming "Litecoin" is what Abe calls the chain: - Stop all processes that might load blocks.
- Run: python -m Abe.admin delete-chain-blocks Litecoin. This should delete all block header information but not the transactions.
- Run Abe with --no-serve --rescan options. This should reinsert the block information.
- Restart stopped processes.
Alternatively, a full reload using a single process will work, if anything will. Try to separate loading from serving web pages, if you haven't already. Only a single process should load blocks until we fix the bugs. |
|
|
|
sudo: no tty present and no askpass program specified
Sorry, try again.
sudo: no tty present and no askpass program specified
Sorry, try again.
sudo: no tty present and no askpass program specified
Sorry, try again.
sudo: 3 incorrect password attempts
[Fri Feb 28 09:37:59 2014] [warn] [client 108.162.221.55] (104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server
[Fri Feb 28 09:37:59 2014] [error] [client 108.162.221.55] Premature end of script headers: abe.fcgi
Trying to use FastCGI as outlined...having issues. Please verify the line that you added to /etc/sudoers. |
|
|
|
Another question, is it possible to set the magic address in the config and not DataStore.py? I would like to parse multiple blocks but have separate databases for them. It seems I can only do this by running separate instances of Abe.
Looks like I might be able to specify the magic when doing 'datadir+=' in the .conf but it wont insert it into the database. You don't need the magic number if you supply the chain name. Magic number is for datadir=/path so Abe can find the chain name given only the magic number found in the block files. When Abe sees a datadir with a previously unknown chain name, it copies the magic number and some other fields to the chain table. On subsequent runs, it does not overwrite these values based on datadir. You would have to update the row or rebuild the database with the modified datadir configuration. Hope this helps. |
|
|
|
Opened /home/nathan/.tekcoin/blk0001.dat
Exception at 183
Failed to catch up {'blkfile_offset': 0, 'blkfile_number': 1, 'chain_id': 8L, 'loader': None, 'dirname': u'/home/nathan/.tekcoin', 'id': 1L}
Traceback (most recent call last):
File "Abe/DataStore.py", line 2637, in catch_up
store.catch_up_dir(dircfg)
File "Abe/DataStore.py", line 2895, in catch_up_dir
store.import_blkdat(dircfg, ds, blkfile['name'])
File "Abe/DataStore.py", line 3025, in import_blkdat
store.import_block(b, chain_ids = chain_ids)
File "Abe/DataStore.py", line 1756, in import_block
raise MerkleRootMismatch(b['hash'], tx_hash_array)
MerkleRootMismatch: Block header Merkle root does not match its transactions. block hash=00000eaa8dd7e902871c444aeb415673b8244c6b7b1742ca398cefa60d87c4c1
Please see: https://bitcointalk.org/index.php?topic=22785.msg750417#msg750417 |
|
|
|
Thanks... So I got to the ppcoin branch, and I merged...but I get a schema version error: Exception: Can not upgrade from schema version Abe37 to Abe35 What happens when you try with a fresh database? |
|
|
|
How does Abe handle network forks?
It chooses the longest chain by cumulative work, like the Bitcoin client. This code could use a test suite, but it handles the 2013 fork, as far as I can tell. Abe keeps orphan blocks and side chains and reuses them if they join the main chain. |
|
|
|
OK, I think i found where to find the address and magic...but now i get the error Using... python -m Abe.abe --config abe-tek.conf --commit-bytes 100000 --no-serve Opened /home/nathan/.tekcoin/blk0001.dat
Exception at 72056512226263178 The block format differs from Bitcoin's. Please see: https://bitcointalk.org/index.php?topic=22785.msg2796317#msg2796317 |
|
|
|
Thank you. Will the newer commits to master be merged into no-statistics? Are there any important changes since 6 months ago? I merge changes from time to time or when people hire me to. As far as I know, no-statistics can handle BTC "mainnet" as-is or with only small fixes from master. |
|
|
|
I'm trying to set up an abe instance for a coin that is primarily proof-of-stake. I see there were mods for NVC and Cash, do I need to add something similar? Any special config required for proof-of stake? I would like to make these chain properties configurable. In the meantime, you can do as I did for NVC and Cash. In the read-me i noticed that the web interface doesn't understand proof of stake transactions. Since this coin is going to be nearly all proof of stake will abe even work? I've removed that bit from README. The block page now does say something about proof-of-stake, example: http://115.29.51.156:7703/chain/CashCoin/b/52705. Thanks. |
|
|
|
I see 985c6346 tries to fix this. Any problem after that commit? |
|
|
|
Hi, I am working on Crapples block explorer and get the following error upon running python -m Abe.abe --config abe.conf --commit-bytes 100000 --no-serve Exception at 103
Failed to catch up {'blkfile_number': 1, 'dirname': u'/root/.crapples', 'chain_id': Decimal('8'), 'id': 2L, 'blkfile_offset': 0}
Traceback (most recent call last):
File "Abe/DataStore.py", line 2422, in catch_up
store.catch_up_dir(dircfg)
File "Abe/DataStore.py", line 2475, in catch_up_dir
store.import_blkdat(dircfg, ds, filename[0])
File "Abe/DataStore.py", line 2599, in import_blkdat
store.import_block(b, chain_ids = chain_ids)
File "Abe/DataStore.py", line 1675, in import_block
raise MerkleRootMismatch(b['hash'], tx_hash_array)
MerkleRootMismatch: Block header Merkle root does not match its transactions. block hash=e3328341fb132cdbd261ba75472d8e85a5db48b987bc8f13ea244fd470040000 Please see: https://bitcointalk.org/index.php?topic=22785.msg750417#msg750417 |
|
|
|
|
Show Posts
