Look into the return latency of the price call

[muneeb-ali]

It seems like the price call takes a while to return. Cryptocracy ran into this issue in his demo.

@jcnelson can you kindly post more info on where the current bottleneck is?

[jcnelson]

It's in pybitcoin's use of /dev/random. I've sent a PR here: stacks-archive/pybitcoin#59.

I can also try to remove pybitcoin from Blockstack Core, which is a long-term goal anyway.

[muneeb-ali]

OK, cool. Looks like circle-ci tests are the current roadblock.

How about we try to get the PR into pybitcoin in the short-term and look into getting away from pybitcoin later?

Also, would love to know more about the dependency i.e., how is the available entropy getting looped into the price call path. If you can point to where BitcoinPrivateKey() is being used in the call flow then I can just pick the logic from there -- thanks!

[jcnelson]

I've updated the rc-0.14.2 so we never call into code that calls pybitcoin's dev_random_entropy() function, which avoids the problem entirely. It also gets rid of the need for rng-tools.

[muneeb-ali]

Awesome! Should we mark this as fixed-in-next-release?

Looks like you switched to keylib in f777a89 which uses urandom. Makes sense.

[jcnelson]

Let me try it on a VM that has exhausted /dev/random to confirm :)

[kantai]

After pushing this to RC14.2, fake private keys are hardcoded, which speeds up the price calculation a lot.

$ time curl -s https://core.blockstack.org/v1/prices/names/muneeb.id > /dev/null
real 0m1.496s

[jcnelson]

This is all fixed in rc-0.14.2.

[jackzampolin]

Closing as fixed! 🎉

[blockstack-devops]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.