On Thu, Feb 27, 2020 at 12:24 AM 4e868df3 <4e868df3@xxxxxxxxx> wrote: > > The VM is reporting the errors. What are the details of the storage stack? What are the Btrfs devices backed by on the host? Physical partitions, or qcow2, or raw files? If they are files, is chattr +C set? And what is the cache method being used for each device? >A power failure or crash is likely to > blame. The host or the VM? The VM can crash, it's fairly safe, I even use unsafe cache on all my VMs (it's way faster) and a force power off on VMs all the time with Btrfs and never have a problem. But, it's a different story if the host crashes. It's a little strange that only two devices are affected. >Scrubs happen monthly (3 weeks ago). No other BTRFS messages in > dmesg to report, just a bunch of sudo audit's. > > $ sudo btrfs scrub status /mnt/raid > UUID: 8c1dea88-fa40-4e6e-a1a1-214ea6bcdb00 > Scrub started: Thu Feb 27 00:16:58 2020 > Status: aborted > Duration: 0:02:03 > Total to scrub: 5.76TiB > Rate: 853.91MiB/s > Error summary: csum=278 > Corrected: 0 > Uncorrectable: 278 > Unverified: 0 Btrfs check might take a while on 5.7T. It depends on how much metadata there is. -- Chris Murphy
