why bts can't recover quickly when crashed?

Started by sudo, February 25, 2016, 01:01:54 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

chryspano

Don't worry guys, once the "barbarians" are here, they'll do all the coding and all the work!

http://www.cavafy.com/poems/content.asp?id=119&cat=1




You only need to have some faith in delutional alt and in buthurt btswildpig!



abit

Quote from: sudo on February 26, 2016, 12:36:12 AM
Quote from: pc on February 25, 2016, 05:26:38 PM
The witness_node keeps all of its internal data in memory. On a regular shutdown, it dumps that internal state to disk. When the node is started it tries to read that internal state from disk. If that works it can continue where it left off, but if it doesn't work it has to re-scan the whole blockchain, which takes some time.

some check point  like every hour or every day set a checkpoint   or made the past blockchain  readonly cached into memory

download oneday data   is much faster than  replaying all the data
Afaik it's under development.
https://github.com/cryptonomex/graphene/issues/499

Funds/votes please? Or you'll like to go on supporting @alt to not develop it or hope someone will develop it for free?
https://cryptofresh.com/u/ags
BitShares committee member: abit
BitShares witness: in.abit

sudo

#4
Quote from: pc on February 25, 2016, 05:26:38 PM
The witness_node keeps all of its internal data in memory. On a regular shutdown, it dumps that internal state to disk. When the node is started it tries to read that internal state from disk. If that works it can continue where it left off, but if it doesn't work it has to re-scan the whole blockchain, which takes some time.

some check point  like every hour or every day set a checkpoint   or made the past blockchain  readonly cached into memory

download oneday data   is much faster than  replaying all the data

pc

The witness_node keeps all of its internal data in memory. On a regular shutdown, it dumps that internal state to disk. When the node is started it tries to read that internal state from disk. If that works it can continue where it left off, but if it doesn't work it has to re-scan the whole blockchain, which takes some time.
Bitcoin - Perspektive oder Risiko? ISBN 978-3-8442-6568-2 [url="http://bitcoin.quisquis.de"]http://bitcoin.quisquis.de[/url]

sudo

if i zip  witness_node_data_dir 
start the witness_node.exe  much faster  sync  up to date then  rescan after witness_node crash
why?

it's also happened in bts1.0

sudo

witness_node.exe quit  suddenly
restart
notice   detect unclean shutdown
then  long time scan
every time must use  CTRL+C to shutdown  then it's safe
it's also bother the exchange  too