Matthew Ahrens

Example output today. Workload is random reads from 4x as many threads as there are CPU's. The kretprobe fires 1/4 as many times as the kprobe. ``` $ sudo bpftrace...

It's great to see progress on being able to specify `maxactive`, thank you! Have y'all given any thought to how bpftrace could retrieve `nmissed`, and print it with a warning...

@danobi Thanks for working on this. What's the status on your kernel patch? It looks like there was some discussion in September/October but I didn't see any conclusion.

@danobi Any update on this?

Sounds like something is causing it to do txg_wait_synced() (push out txg's) very frequently. Given that this is a simple send stream (only one snapshot, no encryption), I'm not sure...

Interesting, so it looks like you are receiving multi-slot dnodes. Have you changed the `dnodesize` property from its default of `legacy`?

The comments seem to indicate that this (txg_wait_synced) should happen if a multi-slot dnode overwrites a single-slot dnode. So I'd think that once you change dnodesize from `auto` (1K) back...

FYI, this worked for me in a simple test. I'm not sure what's different about other configurations that makes it sometimes not work. ``` $sudo zdb -lu /dev/sdc1 | grep...

The theoretical arguments both ways are interesting. However, practical experience can help us weigh the importance of the pros and cons. @GregorKopka and @richardelling, are you using compressed_arc_enabled=0? If so...

@richardelling > relatively few people use L2ARC, while everyone uses ARC. But (we think) almost nobody sets `compressed_arc_enabled=0`. Certainly fewer folks than are using L2ARC. > As for recompressing, AIUI,...