Hello, Thank you for bcache; I've been using it successfully with almost no problems. There's just one issue that keeps coming up for me and I haven't found a resolution. Sometimes a bcache device can be stuck in a registered but unusable state. Here is an example with annotations from the kernel log. 1. Initial state: backing device /dev/sdi operational and bcache running; bcache device is in use (part of a mounted btrfs filesystem). [5157644.106076] bcache: register_bdev() registered backing device sdi [5157644.116702] bcache: bch_cached_dev_attach() Caching sdi as bcache1 on set f38239bc-f726-4844-bf2b-39d400993a34 2. Backing device goes offline from poweroff, disconnect, USB reset, etc. This is not the fault of bcache. [5231620.175549] usb 4-1.2: USB disconnect, device number 39 [5231620.234535] sd 13:0:0:0: [sdi] Synchronizing SCSI cache [5231620.234572] sd 13:0:0:0: [sdi] Synchronize Cache(10) failed: Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK [5231625.230117] bcache: cached_dev_status_update() sdi: device offline for 5 seconds [5231625.230125] bcache: cached_dev_status_update() bdev49: disable I/O request due to backing device offline 3. Backing device comes back online, but this time with a different device name (/dev/sdi now becomes /dev/sdn, in this example). [5231626.926953] usb 4-1.2: new SuperSpeed Plus Gen 2x1 USB device number 44 using xhci_hcd [5231626.956905] usb 4-1.2: New USB device found, idVendor=174c, idProduct=55aa, bcdDevice= 1.00 [5231626.956913] usb 4-1.2: New USB device strings: Mfr=2, Product=3, SerialNumber=1 [5231626.956917] usb 4-1.2: Product: ASM1351 [5231626.956919] usb 4-1.2: Manufacturer: Asmedia [5231626.956922] usb 4-1.2: SerialNumber: 123456789205 [5231627.009822] scsi host18: uas [5231627.010226] scsi 18:0:0:0: Direct-Access ASMT 2135 0 PQ: 0 ANSI: 6 [5231627.011501] sd 18:0:0:0: Attached scsi generic sg8 type 0 [5231627.011681] sd 18:0:0:0: [sdn] Spinning up disk... [5231628.046082] .......ready [5231634.307131] sd 18:0:0:0: [sdn] 23437770752 512-byte logical blocks: (12.0 TB/10.9 TiB) [5231634.307140] sd 18:0:0:0: [sdn] 4096-byte physical blocks [5231634.307263] sd 18:0:0:0: [sdn] Write Protect is off [5231634.307266] sd 18:0:0:0: [sdn] Mode Sense: 43 00 00 00 [5231634.307400] sd 18:0:0:0: [sdn] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA [5231634.307800] sd 18:0:0:0: [sdn] Preferred minimum I/O size 4096 bytes [5231634.307803] sd 18:0:0:0: [sdn] Optimal transfer size 33553920 bytes not a multiple of preferred minimum block size (4096 bytes) [5231634.331809] sd 18:0:0:0: [sdn] Attached SCSI disk 4. Bcache tries to attach the backing device, but it's still attached. [5231634.503820] bcache: register_bdev() registered backing device sdn [5231634.503827] bcache: bch_cached_dev_attach() Tried to attach sdn but duplicate UUID already attached 5. The filesystem has errors. [5233416.742602] BTRFS error (device dm-2): bdev /dev/mapper/backup1 errs: wr 0, rd 1, flush 0, corrupt 0, gen 0 [5233416.742646] BTRFS error (device dm-2): bdev /dev/mapper/backup1 errs: wr 0, rd 2, flush 0, corrupt 0, gen 0 [5233416.745538] BTRFS error (device dm-2): bdev /dev/mapper/backup1 errs: wr 0, rd 3, flush 0, corrupt 0, gen 0 [5233416.747267] BTRFS error (device dm-2): bdev /dev/mapper/backup1 errs: wr 0, rd 4, flush 0, corrupt 0, gen 0 [5233416.747405] BTRFS error (device dm-2): bdev /dev/mapper/backup1 errs: wr 0, rd 5, flush 0, corrupt 0, gen 0 ...etc. Manually stopping the device doesn't work; the "bcache" symlink is a broken link to the device under the old name. $ echo 1 | sudo tee /sys/block/bcache1/bcache/stop tee: /sys/block/bcache1/bcache/stop: No such file or directory 1 $ readlink /sys/block/bcache1/bcache ../../../pci0000:00/0000:00:01.3/0000:01:00.2/0000:02:09.0/0000:07:00.0/usb4/4-1/4-1.2/4-1.2:1.0/host13/target13:0:0/13:0:0:0/block/sdi/bcache At this point, using the backing device via bcache seems impossible without a reboot. * original bcache device is still active * can't stop the bcache device, since the backing device has a new name * the kernel won't give the backing device the old name Is there any way to recover from this situation without a reboot? I'm running the Debian-provided 6.1.0-8-amd64 kernel on Debian sid. I've seen this issue before with older kernels, too. Thank you, Corey
