HP XC System Software Administration Guide Version 3.2

Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0
Number Major Minor RaidDevice State
0 8 1 0 active sync /dev/sda1
1 8 17 1 active sync /dev/sdb1
UUID : eead90a0:35c0bf46:9160b26b:2d754a4d
Events : 0.10
Nagios uses the mdadm command to verify the status of the RAID array.
19.5 Error Reporting
Errors can be reported during the installation of software RAID on a client node. If the installation
software does not find two or more disks, the client node is imaged without software RAID, and
a warning message is sent to the /hptc_cluster/admin/logs/imaging.log log file.
Nagios reports the following types of errors specific to software RAID-1 on the HP XC system:
Fail An active component of the software RAID-1 array was marked as
faulty.
Device Disappeared A previously configured RAID array no longer appears to be
configured.
Degraded arrays
The mdadm command generates this message when it first notices
a degraded drive failure.
“Software RAID Disk Replacement” (page 272) provides information on how to replace a defective
RAID disk and write a new boot block.
19.6 Removing Software RAID from Client Nodes
Removing Software RAID from a client node requires that you reimage and shut down the node.
This example uses three partitions on an IDE disk.
Follow this procedure to remove Software RAID from client nodes:
1. Log in as superuser (root) on the head node.
2. Use the editor of your choice to edit the /etc/systemimager/systemimager.conf file:
a. Locate the following line in the file. The value x represents the software RAID level.
SOFTWARE_RAIDx_NODES = list of nodes
b. Remove the client node from the list of nodes.
c. Save the file and exit.
3. Rerun the cluster_config utility.
4. Remove all the partitions on the second disk:
# mdadm /dev/md1 -r /dev/hdb1
# mdadm /dev/md2 -r /dev/hdb2
# mdadm /dev/md3 -r /dev/hdb3
5. Remove the second disk. Repartition it and remove all drive labels. Although the second
disk will not be touched, it will contain mirrored partitions and conflicting labeled partitions.
6. Reimage the node.
19.5 Error Reporting 229