Rob Norris
Rob Norris
Further testing shows the bug's impact is a little wider: if multiple disks are lost on the same txg causing the pool to suspend, after return they will all re-fault...
There's some good ideas and thinking in here, and I don't have anything to contribute to this bikeshed, so I will not. I did want to just raise a point...
FWIW, mechanically changing `[1]` to `[]` is not necessarily safe, as it changes the result of `sizeof (...)` on that type, so matching changes may need to be made elsewhere.
Duplicate of #16278.
@allanjude @don-brady FYI, this will need a rebase onto #15895 and zdb will then need some rework, as its structure has changed a bit. I'll have a look at it...
Unable to reproduce on 13.2-RELEASE-p11, OpenZFS c98295e. Tried VM with 2x and 4x cores, and 2G and 16G RAM. Typical run: ``` robn@freebsd13:~/zfs $ ./scripts/zfs-tests.sh -Dvxt cp_stress --- Cleanup ---...
I ran the test over and over for a few hours (I forgot about it...), no dice. I set it up to run many thousands of files & threads for...
Oh I'm sorry, my question wasn't clear (also it was more of a drive-by mumble than a really big-brained thought). I more meant, if we know this test is running...
Oof, hard to argue with results. Also yikes, I'll check it out! :sweat_smile:
Unfortunately it's not about where things are accounted to, as such. We don't get to choose that. Linux accounts a certain type of memory through MemAvailable (roughly, "page cache"). That's...