3

So I am trying to figure why geth is such heavy disk I/O at startup, even after the chain is fully synced? The problem is that on a normal non-SSD hard drive it doesn't matter how fast the CPU is, Mist always falls out of sync at least initially and takes forever to get on the same page with geth. This affect is amplified on slower machines to the point I can't ever seem to sync up even after 30 / 40 minutes or more of trying. I had hoped the 1.4.x releases would help, but they do the same thing.enter image description here

  • It's probably writing / updating the state into the database. I'm currently running 9 clients and also noticed a huge I/O bottleneck. – Afr May 10 '16 at 13:43
  • Interesting. I wonder how necessary this level of I/O really is. Potentially it is a huge problem, possibly killing feasibility on mid range hardware, let alone lower end stuff. Hopefully they address it fairly soon for mass adoption's sake. Nobody will wait 3 hours for a slower machine to chill out and run well. – Michael Kilday May 10 '16 at 15:12
  • 1
    We're already working on reducing the I/O – Jeffrey W. May 12 '16 at 14:11
  • I have correlated the Mist sync issues to watching contracts / tokens on non-SSD hard drives, especially on systems with low CPU resources. External links of interest are: github.com/ethereum/mist/issues/425 , reddit.com/r/ethereum/comments/4kkxfp/… – Michael Kilday May 23 '16 at 12:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.