CharlesT423
November 17th, 2014, 03:10 PM
I'm having a problem I can't seem to resolve in regards to an LVM with two PVs: one a physical disk in the server (via a hardware RAID card) and the other through an iSCSI SAN connection. I originally setup the LV on just the physical disk and then a few days ago I added in more space via the iSCSI SAN, which was using /dev/dm-5 as the PV. I did a kernel upgrade last night and rebooted the server, and now I can't get the LV to work. After the reboot the device changed from /dev/dm-5 to /dev/dm-4 (which *shouldn't* matter, but it seems to) and when I run vgchange -a y to scan the volumes I get the following errors:
On the command line:
# vgchange -a y
device-mapper: reload ioctl on failed: Invalid argument
1 logical volume(s) in volume group "sasdrives" now active
2 logical volume(s) in volume group "db1" now active
in dmesg:
[ 422.117201] device-mapper: table: 252:5: linear: dm-linear: Device lookup failed
[ 422.117259] device-mapper: ioctl: error adding target to table
252:5 corresponds to /dev/dm-5 Now the file /dev/dm-5 exists, but it does not show up in multipath -ll or dmsetup ls/dmesetup table. Removing the file doesn't make the error go away. I've updated the filter in /etc/lvm/lvm.conf to exclude dm-5 but that doesn't seem to make a difference. The issue seems to be that /dev/dm-5 doesn't point to an actual device and the kernel won't reload the LVM tables if it can't read that device. Simply adding new drives to the system doesn't help because they become dm-6/7 instead of dm-5.
Any ideas?
On the command line:
# vgchange -a y
device-mapper: reload ioctl on failed: Invalid argument
1 logical volume(s) in volume group "sasdrives" now active
2 logical volume(s) in volume group "db1" now active
in dmesg:
[ 422.117201] device-mapper: table: 252:5: linear: dm-linear: Device lookup failed
[ 422.117259] device-mapper: ioctl: error adding target to table
252:5 corresponds to /dev/dm-5 Now the file /dev/dm-5 exists, but it does not show up in multipath -ll or dmsetup ls/dmesetup table. Removing the file doesn't make the error go away. I've updated the filter in /etc/lvm/lvm.conf to exclude dm-5 but that doesn't seem to make a difference. The issue seems to be that /dev/dm-5 doesn't point to an actual device and the kernel won't reload the LVM tables if it can't read that device. Simply adding new drives to the system doesn't help because they become dm-6/7 instead of dm-5.
Any ideas?