On Sun, Oct 11, 2020 at 03:47:31PM +0200, Reinoud Zandijk wrote: > Hi Paul, > > quick question, is it running in qemu? I stumbled on the fact that the ORDER > of directives to qemu makes a difference! If i append `-snapshot' as last to > qemu ie after the -drive, the drive image gets corrupted anyway(!) since > apparently some writes are done wrong and the result looks quite similar to > you problem. If I prepend the -snapshot before the -drive it doesn't get > corrupted nor touched. > > As for without -snapshot, i haven't dug into that yet
This particular corruption was on real hardware, on a raidframe raid1 over a couple of SATA wd drives. However, I also managed to see corruption in a qemu vm, on with image on local disk, and without snapshot, and vm crashes: http://mail-index.netbsd.org/current-users/2020/05/02/msg038496.html I had a fairly long email debugging session with Jaromir, and he came up with a bunch of patches around wd with broken retry behaviour. I can't recall if all those patches are in current or not. -- Paul Ripke "Great minds discuss ideas, average minds discuss events, small minds discuss people." -- Disputed: Often attributed to Eleanor Roosevelt. 1948.
