On Wed, Oct 12, 2016 at 01:07:46PM +0200, Marcin Mirosław wrote: > W dniu 12.10.2016 o 03:26, Kent Overstreet pisze: > Hi, > [...] > > You're not the only one who's seeing this, I've been chasing this bug for the > > past several days... > > > > Is something that can be done on my side to help with it? > I can add information that tier0 device is a device with short time of > writing requests - it's a HW controller with RAM. Second bling shot is > that using btree-node=256k was problematic some time ago: > http://www.spinics.net/lists/linux-bcache/msg04178.html , maybe this is > somehow related to the btree-node size? It turns out it's not a bcache bug - FUA is buggy... If you want, you could confirm by changing journal_flushes_device() to always return true - journal.h line 85. Gonna try and figure out what the actual FUA bug is tomorrow... -- 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
