3 Replies Latest reply on May 16, 2020 4:25 PM by allenng

    Rpi desktop kit - SSD disappears after a while

    allenng

      I have an SSD connected to my Raspberry Pi 3B+ using the Element14 Pi Desktop kit.  The drive is recognized as `/dev/sda` in Raspbian Buster and correctly mounted from `/etc/fstab` at boot, but then will disappear after running for a while.

       

      I started to suspect insufficient power (despite Element14's assertion that a 2.5A power supply was sufficient to power the rpi+SSD), so I bought a new power supply that provided up to 3.5A (I have yet to confirm that with my Kill-a-Watt).  I rebooted the pi and and drive mounted correctly as usual, but after about 10 hours, `dmesg` showed:

      ```

      [Fri Apr 17 18:45:11 2020] usb 1-1.3: reset high-speed USB device number 5 using dwc_otg

      [Fri Apr 17 18:45:11 2020] sd 0:0:0:0: [sda] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x03 driverbyte=0x00

      [Fri Apr 17 18:45:11 2020] sd 0:0:0:0: [sda] tag#0 CDB: opcode=0x28 28 00 00 0c e0 00 00 00 f0 00

      [Fri Apr 17 18:45:11 2020] print_req_error: I/O error, dev sda, sector 843776

      [Fri Apr 17 18:45:42 2020] sd 0:0:0:0: [sda] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x03 driverbyte=0x00

      [Fri Apr 17 18:45:42 2020] sd 0:0:0:0: [sda] tag#0 CDB: opcode=0x28 28 00 00 0c e0 f0 00 00 10 00

      [Fri Apr 17 18:45:42 2020] print_req_error: I/O error, dev sda, sector 844016

      [Fri Apr 17 18:45:42 2020] usb 1-1.3: USB disconnect, device number 5

      [Fri Apr 17 18:45:42 2020] print_req_error: I/O error, dev sda, sector 105121792

      [Fri Apr 17 18:45:42 2020] Buffer I/O error on dev sda1, logical block 13139968, lost sync page write

      [Fri Apr 17 18:45:42 2020] print_req_error: I/O error, dev sda, sector 843776

      [Fri Apr 17 18:45:42 2020] JBD2: Error -5 detected when updating journal superblock for sda1-8.

      [Fri Apr 17 18:45:42 2020] Aborting journal on device sda1-8.

      [Fri Apr 17 18:45:42 2020] print_req_error: I/O error, dev sda, sector 105121792

      [Fri Apr 17 18:45:42 2020] Buffer I/O error on dev sda1, logical block 13139968, lost sync page write

      [Fri Apr 17 18:45:42 2020] JBD2: Error -5 detected when updating journal superblock for sda1-8.

      [Fri Apr 17 18:45:42 2020] print_req_error: I/O error, dev sda, sector 50333960

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572963: block 6291494: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572961: block 6291494: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda2): ext4_find_entry:1455: inode #131073: comm bash: reading directory lblock 0

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572956: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572967: block 6291494: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda2): ext4_find_entry:1455: inode #131073: comm bash: reading directory lblock 0

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572960: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda2): ext4_find_entry:1455: inode #131073: comm bash: reading directory lblock 0

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572955: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda2): ext4_find_entry:1455: inode #131073: comm bash: reading directory lblock 0

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572951: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572950: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda2): ext4_find_entry:1455: inode #131073: comm bash: reading directory lblock 0

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572965: block 6291494: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] EXT4-fs error (device sda1): __ext4_get_inode_loc:4708: inode #1572952: block 6291493: comm ldconfig: unable to read itable block

      [Fri Apr 17 18:45:42 2020] Buffer I/O error on dev sda2, logical block 15761408, lost sync page write

      [Fri Apr 17 18:45:42 2020] JBD2: Error -5 detected when updating journal superblock for sda2-8.

      [Fri Apr 17 18:45:42 2020] Aborting journal on device sda2-8.

      [Fri Apr 17 18:45:42 2020] Buffer I/O error on dev sda2, logical block 15761408, lost sync page write

      [Fri Apr 17 18:45:42 2020] JBD2: Error -5 detected when updating journal superblock for sda2-8.

      ```

       

      After this, the drive no longer even appears in `lsblk`.  It looks as though something is resetting, but I don't know enough about it to dig in much further.  What is causing the drive to disappear?

       

      -----

       

      Update:

      It looks like this might be a kernel bug, but I haven't figured out what to try next.  I'm hesitant to muck with kernel settings that I don't understand.
      https://unix.stackexchange.com/questions/239782/connection-problem-with-usb3-external-storage-on-linux-uas-driver-proble…

        • Re: Rpi desktop kit - SSD disappears after a while
          colporteur

          Have you try moving the device to another USB port?

           

          This is just a WAG (wild ass guess) test that is related to something that happened to me with a Pi and a bluetooth keyboard/touchpad https://www.amazon.ca/Rii-RT-MWK01-Handheld-Rechargeable-Wireless/dp/B01FC7J6NC  last summer.The device was stable provided it was plugged into one side of the four ports but not the other.

           

          This may be a total waste of time and not bring you any closer to discovery. My apologies for this vague description but at the time I was researching SDR and GNU Radio on a Pi and had little time to troubleshoot an errant keyboard. A quick fix that I discovered was if the bluetooth device dongle was plugged into the different pair of USB ports the problem didn't exist.

           

          I made be way out in left field on this. I don't know how the USB's are arranged on the Pi. Your problem description mirrored something I swore I had seen before. After some thought I recall the small keyboard while sitting on the floor of a living room using the giant screen TV as a monitor.

          1 of 1 people found this helpful