This issue was fixed by "xfs_repair -L". 1) Megaraid SAS (Intel's SATA still unaffected, but time was limited), spinning HDDs. 2) XFS v5 (mkfs.xfs -m crc=1,finobt=1 ...) 3) mount -o logbsize=256k,logbufs=8... (noatime,attr2,inode64,noquota still in both cases)
After few hours starts to xfs failures, with deadlocks, dumps, panics. After few reboots - "xfs_repair -L" mostly good (one OSD formatted). 1 of 3 node without Megaraid still unaffected, but mounted safe now anymore. Same case without "logbsize=256k,logbufs=8" still stable for month now (and working now after repair too). PS hammer from git. -- WBR, Dzianis Kahanovich AKA Denis Kaganovich, http://mahatma.bspu.unibel.by/ _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
