Hi Brian, I'll start from the beginning.. I have a mdadm raid 5 array with 4 SATA disks (2xST4000NC000 and 2xST2000NC000) that were attached to a RocketRaid 2722 controller in an Addonics tower all on the same PMP. Something had a fit in there and knocked the RR driver offline taking out the array when I was adding data to the array. I moved the disks off the array onto individual sata controllers on the system and finished my repairs from there as that setup has some kind of issue when I near 89% usage on the array. I ended up with the xfs log and superblock being screwed up. I was running kernel 3.19.18 and upgraded to 4.6.7 in case there was some unknown bug hiding in there. I did all my repair work on 4.6.7 kernel. I could not get the 4.5 tools to sort things out as the log was messed up (the log head/tail could not be found), and the superblock was messed up (if I remember right it was a metadata error). I tried moving the log to an external setup, but xfs_db would complain about corruption at 0x0/0x200 in the superblock and not let me do this. I upgraded to 4.10 tools with hope but no luck any which way. For some reason with 4.5 I could use the xfsrepair -L option and then re-run with xfsrepair and no -L option and it would do something with the log and superblock and I could mount the FS again. Doing that with 4.10 would not work at all. I ended up downgrading to 4.5 tools again and doing the xfsrepair, xfsrepair -L and xfsrepair and I had success. I have no idea why, but it worked and everything is happy again. I'm on kernel 4.8.17 at the moment and so far so good with the drives nowhere near the highpoint card. I've attached a printout of the superblock that was messed and as it is now as well as one of the xfsrepair straces. For whatever reason, when I went to pull the sb values for this e-mail I had 4.10 installed again from a system update and it kept segfalting.. going back to 4.5 worked. If you need more I can dig through syslog and messages and see what I can pull out for you. Andrew On 05/15/2017 07:59 AM, Brian Foster wrote: > On Sat, May 13, 2017 at 08:04:42PM -0400, Andrew Ryder wrote: >> Hello, >> >> I went through a corruption issue on a v5 filesystem I have and the v4.5 of >> the tools fixed the log but 4.10 would only complain about not being able to >> fix it? >> > > Can you provide details please? > > Brian > >> I can provide some background on the issue if that helps but I though it >> might be worthwhile to mention the above..? >> >> Andrew >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >