Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

rolls out merkle shreds to 100% of testnet slots #32316

Merged
merged 1 commit into from
Jun 29, 2023

Conversation

behzadnouri
Copy link
Contributor

@behzadnouri behzadnouri commented Jun 28, 2023

Problem

Ramping up Merkle shreds rollout as the results look good: #31950 (comment)

Summary of Changes

Rolls out merkle shreds to 100% of testnet slots.

@behzadnouri behzadnouri force-pushed the merkle-shreds-rollout branch from 796ff1b to 7fcaca7 Compare June 28, 2023 18:56
@codecov
Copy link

codecov bot commented Jun 28, 2023

Codecov Report

Merging #32316 (7fcaca7) into master (5e830d2) will increase coverage by 0.0%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##           master   #32316   +/-   ##
=======================================
  Coverage    82.0%    82.0%           
=======================================
  Files         773      773           
  Lines      209659   209659           
=======================================
+ Hits       172097   172108   +11     
+ Misses      37562    37551   -11     

@steviez
Copy link
Contributor

steviez commented Jun 29, 2023

So kind of eye-balling it from your post in #31950 (comment), it seems as if:

  • The repair/data ratio is improved at towards the tail end of the distribution (bottom right of table), and comparable / marginally worse elsewhere.
    • I guess this would seem to indicate that legacy shreds show more variance on this metric?
  • The propagation time (millis/shred) is improved across the distribution (ie element to element for the table) for merkle-shred-slots over legacy-slots

@behzadnouri
Copy link
Contributor Author

behzadnouri commented Jun 29, 2023

  • I guess this would seem to indicate that legacy shreds show more variance on this metric?

Don't know how to interpret that. But since overall repair numbers are pretty low, and worst slots/nodes have improved, I feel ok with that.

@steviez
Copy link
Contributor

steviez commented Jun 29, 2023

Don't know how to interpret that

Might have abused some terminology, but as a concrete example:

99% slots / 75% nodes
    0.0009 legacy
    0.0016 merkle
99% slots / 99% nodes
    0.0455 legacy
    0.0431 merkle

The "best" 75% of nodes do better with legacy, but the entire cluster (or 99% of it that are reporting metrics) do better with merkle. So, a bigger spread for legacy.

But since overall repair numbers are pretty low, and worst slots/nodes have improved, I feel ok with that.

True - I'm probably guilty of splitting hairs here by comparins 0.09% to 0.16%

@behzadnouri behzadnouri merged commit b722ce4 into solana-labs:master Jun 29, 2023
@behzadnouri behzadnouri deleted the merkle-shreds-rollout branch June 29, 2023 18:46
@behzadnouri behzadnouri added the v1.16 PRs that should be backported to v1.16 label Jun 29, 2023
mergify bot pushed a commit that referenced this pull request Jun 29, 2023
mergify bot added a commit that referenced this pull request Jun 29, 2023
…32316) (#32336)

rolls out merkle shreds to 100% of testnet slots (#32316)

(cherry picked from commit b722ce4)

Co-authored-by: behzad nouri <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
v1.16 PRs that should be backported to v1.16
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants