In a previous message, Dave Chinner wrote: > > And this says the filesystem is still mounted. Why? > Good question. As mentioned in the CentOS bug report, this is a freshly Kickstarted CentOS 5.9 system, staged with minimal packages. I'm not running any third party daemons or anything like that. I could wipe the system and re-kickstart it again over and over, and the problem will exist with 5.9 and the 348 kernel. I can also repeat this on any x86 based platform with nothing special, no hardware RAID controllers, just generic SATA or SCSI disks. Doesn't happen with any previous CentOS kernels. Also mentioned, if I take a 5.8 system, which is running correctly, and upgrade the kernel from 308 to 348, the problem shows up with the new kernel. If I downgrade 5.9 to 308, the problem is resolved as well. Also if I use ext3 with md raid1, there is no problem either. That's why in my mind, all roads lead to some interaction between XFS and md raid. Of course I could be off with that assessment. My next email will contain the "sysrq t" output that Ben requested. Thanks! -- Tom _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
