Matthew Ahrens

@deiter You'll have more luck getting this reviewed and integrated over at the OpenZFS github project -- I don't think that the illumos PR's are monitored. I took a quick...

The symptoms seem identical to https://github.com/openzfs/zfs/issues/12001 and https://github.com/openzfs/zfs/issues/12270, which don't use zstd. So I don't see any evidence that this is specific to ZSTD. I'm going to mark zstd-specific comments...

There are a few places that can return EIO in this path. Can you try reproducing with SET_ERROR tracing enabled, `$ echo 512 >/sys/module/zfs/parameters/zfs_flags` (see `__set_error()`) and then gather zfs_dbgmsg...

@Low-power thanks!

@robn those seem like good ideas. I was also wondering how important `zap_trunc()` is. It seems like it would be rare that it's the last blockid that's being freed, and...

Nice catch. @gamanakis are you planning to implement the fix for this, and if so do have a rough timeline? If not I will take it.

> could you explain me why prefetched read in arc_mru_ghost and especially arc_mfu_ghost states promoted buffer to arc_mru instead of arc_mfu? I think a prefetched read in arc_mru_ghost means that...

After re-running the tests, the 20.04 tests pass, and the 18.04 test seem to be timing out on other PR's as well (e.g. https://github.com/openzfs/zfs/pull/14131 this [run](https://github.com/openzfs/zfs/actions/runs/3380983153/jobs/5614376018)).

@ryao Would you like to take another look at this? I've made the suggested changes to spelling and comments.

@ryao Apologies, I made the naming change 2 months ago but didn't properly upload it here. I've updated the PR now. Let me know if there are any places that...