Closed systemcrash closed 5 months ago
Please elaborate. This smells like a Russian thought it and wrote it roughly in English.
You're right, I've made it as a draft, and it existed far too long in this state.
I hope to rework it someday, https://www.reddit.com/r/zfs/comments/ujba3i/does_allocation_overhead_still_exist_with/ looks like a well-worded example to use.
Feel free to send PRs if you're interested, I'd be glad to review them!
Hello :)
The link is generally regarding compression, while the essence of the text in this issue is about write performance. The linked text might be added under a separate heading, so.... let's tackle the issue first.
Does my proposed text capture the essence of what you intended to say?
@systemcrash feel free to review me, thank you https://github.com/openzfs/openzfs-docs/pull/509/files
https://github.com/openzfs/openzfs-docs/blob/55a0d1cf73b564a889a3fb616fddf55be0f33f14/docs/Basic%20Concepts/RAIDZ.rst#write
Please elaborate. This smells like a Russian thought it and wrote it roughly in English.
What is "full stripe width"? If you made an assumption of what "full" means - please define it.
What is "one block write"? Abstract, singular? What on earth does "has a write IOPS of one slowest disk because of that" mean?
"of one slowest disk"? Because of "that" what is "that" here?
If I am to take a guess (which a user should not be doing, because the user reads the documentation in order that they not guess) - were you trying to say something like the following?: