I had enabled bcache on writeback, but once when booting my laptop I got the following error message error on ...: corrupted btree at bucket ... disabling caching Since I was on a plane I just forced bcache to continue by doing echo 1 > bcache/running Amazingly the backing device recovered without problem (the power of btrfs, I even checked with btrfs scrub); probably there was not a lot of dirty data. So I'd like to enable bcache caching again (even if there was an error previously, all my data is backuped to several other computers so I can afford to lose it). I recreated my caching device again with make-bcache -B (after cleaning up the signature so it would accept to run), and attached the device to the backing device again. The problem is that it detect that the backing device was uncleanly cut away from its caching device, and 'state' is inconsistent: Mithrim /sys/block/bcache0/bcache $ sudo bcache-super-show /dev/sda3 sb.magic ok sb.first_sector 8 [match] sb.csum 9694497C61198C4 [match] sb.version 1 [backing device] dev.label (empty) dev.uuid 56e8d6b3-9a68-4df3-8f4a-dfbf535a9575 dev.sectors_per_block 1 dev.sectors_per_bucket 1024 dev.data.first_sector 16 dev.data.cache_mode 1 [writeback] dev.data.cache_state 3 [inconsistent] cset.uuid b67e71bc-8b3d-43c5-9048-c3b182cc77ec Of course this disable all effective caching, as seen in the stats or in the caching device bcache/written which are all 0. So I have two questions: 1) was there a better way to recover from the corrupted btree than forcing the backing device to run without the cache? 2) Is there a way to tell bcache to put the backing device back in a consistent state so that caching get enabled again? [I would like not to format it again via a make-bcache -B, my data is backuped but reinstalling the laptop and transfering the data back is annoying] Thanks! -- Damien Robert http://www.normalesup.org/~robert/pro -- 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
