Join GitHub today
GitHub is home to over 20 million developers working together to host and review code, manage projects, and build software together.
ThreadFlushWalletDB takes a long time with a big wallet.dat #178
Comments
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment
billym2k
Feb 5, 2014
Contributor
What is ThreadFlushWalletDB good for? Does it occur frequently while running dogecoind? Could it result in 100% cpus?
|
What is ThreadFlushWalletDB good for? Does it occur frequently while running dogecoind? Could it result in 100% cpus? |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment
Wack0
Feb 6, 2014
According to coding.md, ThreadFlushWalletDB closes wallet.dat if it hasn't been used for 500ms. Which is fine for a smaller wallet.dat but obviously would cause issues with a big one.
It does occur frequently when running dogecoind. I ended up rebuilding my wallet.dat to stop the lag, but after a week or so it grew to 50MB and dogecoind stopped syncing with the block chain after a while of it running (so payments never got sent). So it looks like there's some bug somewhere, but no idea where: as a temporary solution I can rebuild wallet.dat every week.
(I also am now using the original dogecoind with ThreadFlushWalletDB so I can check debug.log and see how long it takes with a wallet.dat of however much size)
Wack0
commented
Feb 6, 2014
|
According to coding.md, ThreadFlushWalletDB closes wallet.dat if it hasn't been used for 500ms. Which is fine for a smaller wallet.dat but obviously would cause issues with a big one. It does occur frequently when running dogecoind. I ended up rebuilding my wallet.dat to stop the lag, but after a week or so it grew to 50MB and dogecoind stopped syncing with the block chain after a while of it running (so payments never got sent). So it looks like there's some bug somewhere, but no idea where: as a temporary solution I can rebuild wallet.dat every week. (I also am now using the original dogecoind with ThreadFlushWalletDB so I can check debug.log and see how long it takes with a wallet.dat of however much size) |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment
asadhaider
Feb 6, 2014
Contributor
I've been having this issue as well, flushing wallet.dat takes up to 10 seconds sometimes in my wallet and RPC commands seem to hang until the flushing has finished (they're definitely delayed).
This message in the bitcoin development mailing list mentions the issue too:
http://sourceforge.net/mailarchive/message.php?msg_id=31482185
|
I've been having this issue as well, flushing wallet.dat takes up to 10 seconds sometimes in my wallet and RPC commands seem to hang until the flushing has finished (they're definitely delayed). This message in the bitcoin development mailing list mentions the issue too: |
ummjackson
self-assigned this
Feb 9, 2014
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment|
does this problem still exist in Dogecoin Core 1.7? |
leofidus
added
the
enhancement
label
Apr 15, 2014
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment
|
Don't think we'll know until heavy users start with the 1.7 beta... |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
Show comment Hide comment|
Any updates now that more people have large testnet wallets? |
Wack0 commentedJan 30, 2014
I have a 400MB wallet.dat and ThreadFlushWalletDB takes upwards of 60 seconds every time. I've had to compile my own dogecoind without ThreadFlushWalletDB running specifically because this was lagging transactions (even with 100 rpcthreads).
Hopefully this can be fixed.