[13:20] <hallyn> hm, having a weird corruption issue on 4.13 kernel on artful.  Create an empty file (truncate -s 30G), qemu-nbd attach it, create pv/vg/lv and thinpool-provision;  setup a slew of filesystems;  umount, disconnect, and ...   corruption
[13:20] <hallyn> serge@lxd1:~/sandbox$ sudo qemu-nbd -f raw -c /dev/nbd0 /var/lib/lxd/disks/disk2.img
[13:20] <hallyn> Failed to set NBD socket
[13:20] <hallyn> Disconnect client, due to: End of file
[13:21] <hallyn> ohei.  nbd1 doesn't complain though.  
[13:21] <hallyn> (on zesty this all worked fine, exact same recipe, worked many times on zesty, failed 3x on artful)
[13:22] <hallyn> ok so let's try a reboot - though that didn't work last night, but there could always be >1 issues
[13:23] <cpaelzer> hi hallyn: you point to 4.13 - did you already have a chance to run the same with the zesty kernel but the artful lvm/nbd tools?
[13:35] <hallyn> cpaelzer: no, i haven't.  i was kind of constrained in where i coudl run for the results i wanted, but i guess for reproducing this i can be anywhere.
[13:36] <hallyn> well this is frustrating :)
[13:37] <cpaelzer> hallyn: still corrupted?
[13:37] <cpaelzer> hallyn: do you have your "reate pv/vg/lv and thinpool-provision;  setup a slew of filesystems" in a way you could share it to cross check over here?
[13:38] <hallyn> cpaelzer: yeah, mostly - it's kindof a general tool so there's extra cruft, but it's mainly using https://github.com/atom-deps/lpack/blob/master/setup_lvm.sh  and https://github.com/atom-deps/lpack/blob/master/lpack_unpack.sh
[13:39] <hallyn> TBh kernel commit 7a362ea96d0df873397be04f4556e92f7e37c5ec looks suspect: "nbd: clear disconnected on reconnect"
[13:40] <hallyn> All right let me (1) setup a zesty vm and reproduce there, (2) do-release-upgrade -d there but not reboot, then finally reboot into artful kernel;  and see which of those three work
[13:42] <hallyn> but imma do that at the coffee shop, back in a afew
[13:53] <cpaelzer> I tried with losetup but things worked for me (slightly reduced e.g. no snapshotting)
[13:54] <cpaelzer> interested what you'll see in your test in the zesty vm
[13:58] <hallyn> note this is using nbd not losetup
[14:00] <hallyn> lol maybe https://lkml.org/lkml/2017/7/21/294
[15:03] <hallyn> cpaelzer: weeeeeel crap.  cannot reproduce on vm on my own host.
[15:24] <hallyn> oh, well, maybe it's a race.  now it's acting funky.