On 2019/7/29 下午10:46, Swâmi Petaramesh wrote: > Le 29/07/2019 à 16:40, Qu Wenruo a écrit : >> At least from what I see in this transid error, unless you ruled out the >> possibility of bad disk firmware and LVM/LUKS, it's hard to say it's >> btrfs causing the problem. > > Well it would be such a coincidence that a machine that has been 100% > stable since 2014 suddently crashes 2 filesystems on 2 different disks > justy after upgrading to 5.2 and both things would be unrelated. > > Can happen, but the chances... If I understand your two threads correctly, one is this transid error, the other one is tree-checker, which is completely different from this one. The tree-checker one is mostly caused by older fs, and as you mentioned, reverting to 5.1 solves that problem. This transid won't be resolved whatever kernel version you use, it's a real corruption in extent tree, caused by incorrect metadata CoW. So they are two different problems, and this transid error can be completely unrelated to 5.2 kernel. Thanks, Qu > >> In fact, we have a more experienced sysadmin, Zygo, sharing his >> experience of bad *HARDWARE* causing various Flush/FUA problem, which is >> not easy to hit in normal use case, but only after power loss. > >> So for your transid error, it's really hard to pin down the cause, >> unless you have deployed hundreds btrfs... > > For the record I am a professional sysadmin with 20+ years Linux > experience and I have deployed BTRFS on dozens of systems (OK maybe not > hundreds, not sure)...> > Kind regards. > > ॐ >
