Description of problem: On my new, fast machine, the mdadm test very occasionally fails: PASS: test-inspect-fstab.sh PASS: test-inspect-fstab-md.sh PASS: test-list-filesystems.sh PASS: test-list-md-devices.sh libguestfs: error: md_detail: mdadm: md device /dev/md125 does not appear to be active. FAIL: test-mdadm.sh The error message comes from mdadm's '-D' (detail) subcommand when the device in question returns ENODEV. This is the sequence of events (dev == "/dev/md125"): int fd = open(dev, O_RDONLY); # successful vers = md_get_version(fd); # successful, and vers >= 9000 if (ioctl(fd, GET_ARRAY_INFO, &array)<0) { if (errno == ENODEV) ... error message is printed here ... Version-Release number of selected component (if applicable): libguestfs 1.17.31 mdadm-3.2.3-7.fc17.x86_64 How reproducible: infrequently Steps to Reproduce: 1. make -C tests/md check
I think this may have been related to the guest kernel / mdadm utility used in the appliance. I cannot reproduce this. Can we close it?
This is a real old bug, and I also cannot reproduce it - closed.