Automount of large partitions not working

    Greetings,

    i'm new to kodi and libreelec and i have trouble with the automount functionality. I have a raspi 3 with libreelec and kodi 18 (milhouse testbuilds). These working fine since build 1023. I have a raidsonic RAID5 external hdd case with 4x3TB (9TB as RAID 5). In build 1022 the drive is correctly attached but after update to 1023 not. I tried the newest build with no effort. I did some tests and found out, that partitions greater 8.399.999 MB are not mounted on boot. I can mount them manually and then they work as expected but automount did not work anymore.

    How das automount work on libreelec and are there any config options or restrictions that id dont know ?

    Greetings Deepjungle

    Auto mounting is done by udevil, check out /usr/lib/udev/rules.d/95-udevil-mount.rules for how partitions are matched. Size shouldn't matter, at least not regarding disks :)

    Edited once, last by escalade (January 2, 2017 at 4:40 PM).

    I compared the working and the not working rules. Both are identical.

    Edited once, last by Deepjungle (January 2, 2017 at 4:59 PM).

    These are the dmesg logs:

    1022 dmesg log
    KSJO

    1031 dmesg log
    MWYT

    Actually i cant test the manual command, but when i can then i have to use only this one right

    Code
    /usr/bin/systemctl restart udevil-mount@/dev/%k.service

    Edited once, last by Deepjungle (January 2, 2017 at 6:19 PM).

    But why does it work with less than 8.388.240 MB without any problems. I dont know the exact value when it is overturning but it must be between 8.388.240 MB (working) 8.399.999 MB (not working).

    Quote from Deepjungle


    But why does it work with less than 8.388.240 MB without any problems. I dont know the exact value when it is overturning but it must be between 8.388.240 MB (working) 8.399.999 MB (not working).


    The device size is detected different (and you swapped the two log files ;) :(

    1022 dmesg log:

    Code
    [ 9.840137] sd 0:0:0:0: [sda] 17581277184 512-byte logical blocks: (9.00 TB/8.19 TiB)

    1031 dmesg log:

    Code
    [ 9.733719] sd 0:0:0:0: [sda] 23441768448 512-byte logical blocks: (12.0 TB/10.9 TiB)

    But I have currently no idea what is causing this.

    No problem with the device drive. These are two different raspi 3 one with 12TB and the other one with 9TB Raid. ;)
    [hr]

    Quote from escalade


    You have to replace %k with your partition. Just a guess, but perhaps the disk is not quite ready somehow at the time when udevil runs.

    I tried the command on 1022 and 1023

    Code
    /usr/bin/systemctl restart udevil-mount@/dev/sda2.service

    On build 1022 it works fine on 1023 not. I made a sha1 checksum compare of the udev rules on both builds... no difference

    Here the blkid Logs:

    On 1023 the sda2 partition gets not recognized.

    Greetings.

    Edited once, last by Deepjungle (January 3, 2017 at 9:45 AM).

    The udev rule is simply to launch the udevil mount command when the disk is plugged. Then udevil tells the kernel to mount the drive. If it can't then it's a kernel issue, nothing to do with user space. Looks like the kernel version is different, 4.8.2 vs 4.8.6. As it's a milhouse build you should report it in his thread, he will know if there's been a change in kernel configuration or what not. I'd give the latest LE alpha a try as well, it has 4.9 kernel.

    Edited once, last by escalade (January 3, 2017 at 10:09 AM).

    K i will wait for milhouse. I posted the problem in his thread before. i thought i could find the solution by myself but i think im lost with kernel problems.

    thanks for your explanations and help.