Re: suspend/resume versus OpenZFS on USB

On 2020-09-04 01:42, Graham Perrin wrote:
> This week for the first time I toyed with OpenZFS on a USB device: a 
> mobile hard disk drive connected to the dock of an HP EliteBook 8570p.
> 
> A light test, with the pool imported but not writing to the dataset at 
> suspend time.
> 
> At resume time (22:31), the device was still physically connected but 
> the pool suffered an I/O failure (and the keyboard and trackball on USB 
> were unusable).
> 
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11239]: vdev state changed, 
> pool_guid=$8076233369858608335 vdev_guid=$13893535540375859253
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11243]: pool I/O failure, 
> zpool=$Transcend error=$28
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11247]: pool I/O failure, 
> zpool=$Transcend error=$28
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11251]: pool I/O failure, 
> zpool=$Transcend error=$28
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11255]: pool I/O failure, 
> zpool=$Transcend error=$28
> Sep  3 22:31:03 momh167-gjp4-8570p ZFS[11259]: catastrophic pool I/O 
> failure, zpool=$Transcend
> 
> I cleared pool errors (after which the keyboard and trackball became 
> usable), exported the pool, physically disconnected the drive, restarted 
> the OS, reconnected, imported and – luckily – cleared the remaining 
> metadata errors through a scrub.
> 
> Please, might something be done to improve suspend/resume reliability 
> with storage on USB?
> 
> Might I/O failures be less likely if I connect the drive to the notebook 
> instead of its dock?
> 
> TIA
> 
> ----
> 
> More from /var/log/messages at https://pastebin.com/CqRYbFZm
> 
> Other relevant info below.
> 
> root_at_momh167-gjp4-8570p:~ # date ; uptime ; uname -v
> Thu Sep  3 22:45:07 BST 2020
> 10:45PM  up 6 mins, 6 users, load averages: 0.20, 0.27, 0.13
> FreeBSD 13.0-CURRENT #63 r364768: Tue Aug 25 20:08:23 BST 2020 
> root_at_momh167-gjp4-8570p:/usr/obj/usr/src/amd64.amd64/sys/GENERIC-NODEBUG
> root_at_momh167-gjp4-8570p:~ # zpool status -v
>    pool: copperbowl
>   state: ONLINE
> status: Some supported features are not enabled on the pool. The pool can
>          still be used, but some features are unavailable.
> action: Enable all features using 'zpool upgrade'. Once this is done,
>          the pool may no longer be accessible by software that does not 
> support
>          the features. See zpool-features(5) for details.
>    scan: scrub repaired 0B in 01:39:31 with 0 errors on Thu Sep  3 
> 01:12:21 2020
> config:
> 
>          NAME          STATE     READ WRITE CKSUM
>          copperbowl    ONLINE       0     0     0
>            ada0p4.eli  ONLINE       0     0     0
> 
> errors: No known data errors
> root_at_momh167-gjp4-8570p:~ # zpool import Transcend && zfs load-key 
> Transcend/VirtualBox
> Enter passphrase for 'Transcend/VirtualBox':
> root_at_momh167-gjp4-8570p:~ # zpool status -v Transcend
>    pool: Transcend
>   state: ONLINE
> status: One or more devices has experienced an error resulting in data
>          corruption.  Applications may be affected.
> action: Restore the file in question if possible.  Otherwise restore the
>          entire pool from backup.
>     see: https://zfsonlinux.org/msg/ZFS-8000-8A
>    scan: scrub repaired 0B in 00:15:33 with 0 errors on Wed Sep  2 
> 22:38:56 2020
> config:
> 
>          NAME        STATE     READ WRITE CKSUM
>          Transcend   ONLINE       0     0     0
>            da0p1     ONLINE       0     0     0
> 
> errors: Permanent errors have been detected in the following files:
> 
>          <metadata>:<0x0>
>          <metadata>:<0x3d>
> root_at_momh167-gjp4-8570p:~ # zpool scrub Transcend
> root_at_momh167-gjp4-8570p:~ # zfs version
> zfs-0.8.0-1
> zfs-kmod-0.8.0-1
> root_at_momh167-gjp4-8570p:~ # zpool status -v Transcend
>    pool: Transcend
>   state: ONLINE
>    scan: scrub repaired 0B in 00:17:53 with 0 errors on Thu Sep  3 
> 23:03:57 2020
> config:
> 
>          NAME        STATE     READ WRITE CKSUM
>          Transcend   ONLINE       0     0     0
>            da0p1     ONLINE       0     0     0
> 
> errors: No known data errors
> root_at_momh167-gjp4-8570p:~ # zfs get 
> compression,compressratio,encryption,used,referenced,mountpoint 
> Transcend Transcend/VirtualBox
> NAME                  PROPERTY       VALUE SOURCE
> Transcend             compression    zstd                      local
> Transcend             compressratio  1.70x                     -
> Transcend             encryption     off default
> Transcend             used           76.4G                     -
> Transcend             referenced     76.4G                     -
> Transcend             mountpoint     /Volumes/t500             local
> Transcend/VirtualBox  compression    zstd inherited from Transcend
> Transcend/VirtualBox  compressratio  1.00x                     -
> Transcend/VirtualBox  encryption     aes-256-gcm               -
> Transcend/VirtualBox  used           200K                      -
> Transcend/VirtualBox  referenced     200K                      -
> Transcend/VirtualBox  mountpoint     /Volumes/t500/VirtualBox inherited 
> from Transcend
> root_at_momh167-gjp4-8570p:~ # ls -hl 
> /Volumes/t500/VirtualBox/Windows/Windows.vdi
> -rw-------  1 grahamperrin  grahamperrin    69G Sep  3 16:58 
> /Volumes/t500/VirtualBox/Windows/Windows.vdi
> root_at_momh167-gjp4-8570p:~ # du -hs 
> /Volumes/t500/VirtualBox/Windows/Windows.vdi
>   36G    /Volumes/t500/VirtualBox/Windows/Windows.vdi
> root_at_momh167-gjp4-8570p:~ #
> 

We need output from "procstat -akk" to see where ZFS/USB is hanging.

--HPS