On Thu, Aug 25, 2016 at 12:13:53PM +0200, Marcin Mirosław wrote: > Hi! > > 1. > # ./bcache format --compression_type=lz4 --error_action=readonly > --tier=0 /dev/system10/bcache --tier=1 /dev/sdd1 > /dev/system10/bcache contains a bcache filesystem > Proceed anyway? (y,n) y > /dev/sdd1 contains a bcache filesystem > Proceed anyway? (y,n) y > UUID: ab597c6a-7394-41ff-9138-e66c5722bc9d > Set UUID: 311a03a9-3646-40a8-935f-10030ee75b25 > version: 6 > nbuckets: 22288 > block_size: 1 > bucket_size: 1024 > nr_in_set: 2 > nr_this_dev: 0 > first_bucket: 3 > UUID: 2ac11eef-6d66-4ca0-bd6e-17ac83c6942a > Set UUID: 311a03a9-3646-40a8-935f-10030ee75b25 > version: 6 > nbuckets: 40960 > block_size: 1 > bucket_size: 1024 > nr_in_set: 2 > nr_this_dev: 1 > first_bucket: 3 > > > # mount -o noatime -t bcache /dev/system10/bcache:/dev/sdd1 /mnt/test > # dd if=/dev/urandom of=/mnt/test/randomdata bs=1M count=1000 > # md5sum /mnt/test/randomdata > a7d2712c673d891d9ba50f2f7157c091 /mnt/test/randomdata > # cat /sys/fs/bcache/311a03a9-3646-40a8-935f-10030ee75b25/tiering_percent > 10 > # echo 1 > > /sys/fs/bcache/311a03a9-3646-40a8-935f-10030ee75b25/tiering_percent ; > sleep 5 ; umount /mnt/test > > Now I'm getting in dmesg: > Aug 25 11:55:04 localhost kernel: [ 1366.385581] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.385600] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.390298] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.391076] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.391098] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.391111] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.417244] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): IO error: read only > Aug 25 11:55:04 localhost kernel: [ 1366.656319] bcache > (311a03a9-3646-40a8-935f-10030ee75b25): stopped > > Now: > # mount -o noatime -t bcache /dev/system10/bcache:/dev/sdd1 /mnt/test > # md5sum /mnt/test/randomdata > a7d2712c673d891d9ba50f2f7157c091 /mnt/test/randomdata > > So it is ok. But logs in dmesg are a little scary. > I had data corruption in similar situation like this but I can't > reproduce it. > Btw, should be bcachefs immune of power reset while IO activity? Yes, it definitely should be. Do id you see in the log what caused it to go RO? Or did you do that via sysfs? > > 2. Unsuccessful try of unregister device: > # echo 1 > > /sys/fs/bcache/311a03a9-3646-40a8-935f-10030ee75b25/cache1/unregister > > and console hangs. I have acces to files from other console. > In dmesg is: > Aug 25 12:07:48 localhost kernel: [ 2130.432099] INFO: task bash:19379 > blocked for more than 30 seconds. > Aug 25 12:07:48 localhost kernel: [ 2130.432105] Tainted: P > O 4.7.0-bcache+ #5 > Aug 25 12:07:48 localhost kernel: [ 2130.432108] "echo 0 > > /proc/sys/kernel/hung_task_timeout_secs" disables this message. > Aug 25 12:07:48 localhost kernel: [ 2130.432111] bash D > ffff88014857bc88 0 19379 19376 0x00000000 > Aug 25 12:07:48 localhost kernel: [ 2130.432118] ffff88014857bc88 > 0000000000000000 ffff88014a1a1b40 ffff8800c8f78000 > Aug 25 12:07:48 localhost kernel: [ 2130.432124] ffff88014857bcb0 > ffff88014857c000 ffffffffc0785244 ffff8800c8f78000 > Aug 25 12:07:48 localhost kernel: [ 2130.432129] 00000000ffffffff > ffffffffc0785248 ffff88014857bca0 ffffffff8156fcea > Aug 25 12:07:48 localhost kernel: [ 2130.432135] Call Trace: > Aug 25 12:07:48 localhost kernel: [ 2130.432147] [<ffffffff8156fcea>] > schedule+0x3a/0x90 > Aug 25 12:07:48 localhost kernel: [ 2130.432152] [<ffffffff81570163>] > schedule_preempt_disabled+0x13/0x20 > Aug 25 12:07:48 localhost kernel: [ 2130.432155] [<ffffffff81571c3b>] > __mutex_lock_slowpath+0x9b/0x140 > Aug 25 12:07:48 localhost kernel: [ 2130.432159] [<ffffffff81571cf2>] > mutex_lock+0x12/0x30 > Aug 25 12:07:48 localhost kernel: [ 2130.432202] [<ffffffffc075f8be>] > bch_cache_remove+0x1e/0xe0 [bcache] > Aug 25 12:07:48 localhost kernel: [ 2130.432218] [<ffffffffc0762605>] > __bch_cache_store+0x245/0x650 [bcache] > Aug 25 12:07:48 localhost kernel: [ 2130.432234] [<ffffffffc0762a44>] > bch_cache_store+0x34/0x50 [bcache] > Aug 25 12:07:48 localhost kernel: [ 2130.432238] [<ffffffff81206272>] > sysfs_kf_write+0x32/0x40 > Aug 25 12:07:48 localhost kernel: [ 2130.432240] [<ffffffff812057f3>] > kernfs_fop_write+0x113/0x190 > Aug 25 12:07:48 localhost kernel: [ 2130.432243] [<ffffffff8118e5c2>] > __vfs_write+0x32/0x150 > Aug 25 12:07:48 localhost kernel: [ 2130.432247] [<ffffffff812e8a33>] ? > __this_cpu_preempt_check+0x13/0x20 > Aug 25 12:07:48 localhost kernel: [ 2130.432251] [<ffffffff8109f201>] ? > update_fast_ctr+0x41/0x70 > Aug 25 12:07:48 localhost kernel: [ 2130.432253] [<ffffffff8109f262>] ? > percpu_down_read+0x12/0x50 > Aug 25 12:07:48 localhost kernel: [ 2130.432256] [<ffffffff8118f8c3>] > vfs_write+0xb3/0x1b0 > Aug 25 12:07:48 localhost kernel: [ 2130.432258] [<ffffffff81190ce0>] > SyS_write+0x50/0xc0 > Aug 25 12:07:48 localhost kernel: [ 2130.432261] [<ffffffff811adbde>] ? > __close_fd+0x9e/0xc0 > Aug 25 12:07:48 localhost kernel: [ 2130.432264] [<ffffffff8157431f>] > entry_SYSCALL_64_fastpath+0x17/0x93 > > > > iostat show: > # iostat -d 1 3 /dev/sdd1 /dev/mapper/system10-bcache > Linux 4.7.0-bcache+ (marcinm) 25.08.2016 _x86_64_ (4 CPU) > > Device: tps kB_read/s kB_wrtn/s kB_read kB_wrtn > sdd1 0,90 2,40 56,36 5920 138772 > dm-14 21,32 919,34 423,55 2263587 1042859 > > Device: tps kB_read/s kB_wrtn/s kB_read kB_wrtn > sdd1 1,00 0,00 64,00 0 64 > dm-14 2,00 64,00 0,50 64 0 > > Device: tps kB_read/s kB_wrtn/s kB_read kB_wrtn > sdd1 1,00 0,00 64,00 0 64 > dm-14 2,00 64,00 0,50 64 0 > > > Thanks, > Marcin > > -- > To unsubscribe from this list: send the line "unsubscribe linux-bcache" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-bcache" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
