external HDD not detected/mounting v10.95.1 on PI3

Hey all, since the described issue has also plagued me since the release of LE 11 I would like to share my observations too.

By my experience the HDDs are working fine as long as you don't perform any kind of write/delete action using the built in file manager. Once I do that (mostly to delete obsolete media files) and disconnect the drive I have to perform an error check on my Windows PC before it is recognized by LE again.

Here's a log documenting it:

https://paste.libreelec.tv/hot-bug.log

Apr 09 16:31:24.787715 LibreELEC kernel: usb 2-3.3: USB disconnect, device number 5
Apr 09 16:31:24.788413 LibreELEC kernel: device offline error, dev sdb, sector 6324224 op 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 2
Apr 09 16:31:24.788460 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 786432, lost sync page write
Apr 09 16:31:24.788501 LibreELEC kernel: ntfs3: sdb2: ino=3, ntfs_set_state failed, -5.
Apr 09 16:31:24.788542 LibreELEC kernel: ntfs3: sdb2: Mark volume as dirty due to NTFS errors
Apr 09 16:31:24.788582 LibreELEC kernel: device offline error, dev sdb, sector 6324224 op 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 2
Apr 09 16:31:24.788622 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 786432, lost sync page write
Apr 09 16:31:24.788661 LibreELEC kernel: ntfs3: sdb2: ino=3, ntfs_set_state failed, -5.
Apr 09 16:31:24.788699 LibreELEC kernel: device offline error, dev sdb, sector 6324224 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.788739 LibreELEC kernel: device offline error, dev sdb, sector 62336 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.788778 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 3696, lost async page write
Apr 09 16:31:24.788818 LibreELEC kernel: device offline error, dev sdb, sector 5611840 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.788858 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 697384, lost async page write
Apr 09 16:31:24.788905 LibreELEC kernel: device offline error, dev sdb, sector 6318192 op 0x1:(WRITE) flags 0x0 phys_seg 3 prio class 2
Apr 09 16:31:24.788947 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 785678, lost async page write
Apr 09 16:31:24.788986 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 785679, lost async page write
Apr 09 16:31:24.789027 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 785680, lost async page write
Apr 09 16:31:24.789068 LibreELEC kernel: device offline error, dev sdb, sector 6324232 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.789106 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 786433, lost async page write
Apr 09 16:31:24.789147 LibreELEC kernel: device offline error, dev sdb, sector 6327864 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.789187 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 786887, lost async page write
Apr 09 16:31:24.789227 LibreELEC kernel: device offline error, dev sdb, sector 6324224 op 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 2
Apr 09 16:31:24.789267 LibreELEC kernel: Buffer I/O error on dev sdb2, logical block 786432, lost sync page write
Apr 09 16:31:24.789306 LibreELEC kernel: ntfs3: sdb2: ino=3, ntfs3_write_inode failed, -5.
Apr 09 16:31:24.789352 LibreELEC kernel: device offline error, dev sdb, sector 6324224 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
Apr 09 16:31:24.871600 LibreELEC systemd[1]: Stopping udevil-mount@-dev-sdb2.service...
Apr 09 16:31:24.877567 LibreELEC udevil[1532]: udevil: success running umount as current user
Apr 09 16:31:24.879382 LibreELEC systemd[1]: var-media-extern\x20HD\x2011.mount: Deactivated successfully.
Apr 09 16:31:24.888422 LibreELEC systemd[1]: udevil-mount@-dev-sdb2.service: Deactivated successfully.
Apr 09 16:31:24.888739 LibreELEC systemd[1]: Stopped udevil-mount@-dev-sdb2.service.

Apr 09 16:32:19.497713 LibreELEC kernel: ntfs3: sdb2: It is recommened to use chkdsk.
Apr 09 16:32:19.590493 LibreELEC udevil[1605]: mount: /var/media/extern HD 11: wrong fs type, bad option, bad superblock on /dev/sdb2, missing codepage or helper program, or other error.
Apr 09 16:32:19.590493 LibreELEC udevil[1605]:        dmesg(1) may have more information after failed mount system call.
Apr 09 16:32:19.590924 LibreELEC kernel: ntfs3: sdb2: volume is dirty and "force" flag is not set!
Apr 09 16:32:19.613525 LibreELEC systemd[1]: Finished udevil-mount@-dev-sdb2.service.

Is there really no way that LE could handle things like before? Unfortunately formatting the HDDs with a different file system is no option to me ... Thanks for any kind of support!

EDIT before even posting: Here's another try safely removing the device I deleted a file from before switching it off and on again (the faulty sdb2 drive still being active):

https://paste.libreelec.tv/quick-husky.log

This time the drive I deleted a file from (sde2) is recognized upon next startup. And just to prove my point I switched off the drives whithout safely removing them first and voila sde2 doesn't get recognized any more (the follwoing log contains the data from quick husky too):

https://paste.libreelec.tv/desired-chipmunk.log

So maybe this is the resolution to our problems? There might have been a change between LE 11 and 12 though as I faintly remember trying the safe removal before without success.

Hy I try with all versions 11, 12 and same problem

The only solution I find is go back to 10.0.0.4

Think i encounter the same problem now. The external 4TB SSD works perfectly fine on coreelec, even it reports in dmesg that it is flagged "dirty". I guess thats just what happens when the system goes a lot in standby. For me thats unfortunately a killer. Any solution to that? Any way to "repair" the drive on libreelec? If not, guess i'll go with raspbian and install kodi and see if that works.

edit: googling around a bit, i noticed it seems like that mounting RW is blocked by default when the drive is dirty, but RO always works. Thats ok for me (at least for now). mount /dev/sda1 /media -o ro

Is there any way to tell libreelec to mount this device in RO? Or tell libreelec to use -o force? its a movie ssd, i dont care about broken/loss of data, it just has to work.

There are no (reliable) tools for clearing dirty NTFS filesystems, only on Windows.

thats what i was afraid about

any idea how i can configure libreelec to mount that device in RO or with -o force? i assume thats what coreelec does

CE uses NTFS-3G to support NTFS devices, because the older (Amlogic BSP) kernels they use do not have in-kernel drivers. LE devices are not being mounted the same way, and thus experience different issues.

awesome, thanks!

edit: somehow its not working, guess i do something wrong. i changed nano /storage/.config/udevil.conf from default_options_ntfs = nosuid, noexec, nodev, noatime, fmask=0133, uid=$UID, gid=$GID to default_options_ntfs = force, nosuid, noexec, nodev, noatime, fmask=0133, uid=$UID, gid=$GID and tried a mount.ntfs /dev/sda1 /media, but i get the same error.

when i mount it manually with force i get: /dev/sda1 on /var/media type ntfs3 (rw,relatime,uid=0,gid=0,force,iocharset=utf8)

Automount is done via udevil. To test use:

udevil mount /dev/sda1

thanks, very useful information.

udevil: denied 90: option 'force' is not an allowed option
i assume i mis-configured it

changed force to "ro" and now it works at least for reading, but would be awesome if read-write would also somehow work but not sure if this is an option or some other special flag that must be passed to mount

Likely you have to create an allowed_options_ntfs. See the "allowed_options" section.

thanks, that made it