HP Insight Control Software for Linux User Manual

Page 233

Advertising
background image

Corrective actions

Cause/Symptom

the partition wizard, the deployment fails because the
partitions are not sized as they should be.

In this instance, capture the multi-partition image, deploy
it to a single partition, capture that image, then redeploy
the image using the number of required partitions.

Perform the appropriate action:

Captured Image could fail to deploy because of duplicate
labels or existing partition information

Duplicate labels could cause a captured image deployment
failure. A duplicate label can cause the deployed OS to

1.

Diagnose

→Boot to Linux Rescue Mode to access the

RAM Disk environment on the target system.

2.

Use the following command to clear the first 5120 bytes
of the disk to remove any partition or existing master
boot record (MBR) information:

select an incorrect disk to mount or use during the boot
processes. Duplicate labels occur when a disk label from

# dd if=/dev/null of=disk count=1 bs=5120

the image already exists on more than one disk on the
target system.

3.

Rewrite the disk's partition table to remove any existing
labels.

Also, if a disk was previously used as part of an LVM or
RAID set, the partition information could interfere with the
partitioning code during deployment.

Modify the /opt/mx/icle/icle.properties file by
adding the following parameter, where sec is the timeout
value in seconds:

Timeout occurs during Linux image deployment

A timeout can occur when it takes an unusually long time
to prepare a disk for imaging. The default timeout is 30

RAMDISK_DISK_CONFIGURATION_TIMEOUT=sec

minutes, but in rare cases, such as with a very large disk
partition, the preparation might exceed that limit. When

For example, a value of 3600 specifies a timeout of one
hour. You can specify whatever value you believe is

this happens, you can manually increase the timeout value
for this operation.

appropriate, but keep in mind, this increase should only
be needed in unusual circumstances.

HP suggests that you do not create the swap partition in
an LVM volume, and that you do not use the Partition
Wizard for creating or renaming the swap partition.

Cannot create or rename swap partition

The Partition Wizard does not allow the creation or
renaming of a swap partition. The Partition Wizard insists
that you specify a mount point, which swap partitions lack.

See the entry for Bare metal discovery returns "There are
no MAC addresses registered for this system." in

“Troubleshooting discovery problems” (page 213)

Deployment operation returns "There are no MAC
addresses registered for this system."

Use the console to examine the /tmp/rct-debug.log
file.

Deployment operation does not complete.

Wipe the disk that will receive the image and rerun the
deployment operation.

Deployment operation fails.

This pertains to deploying a single Linux image to a server
running Software RAID or LVM volume.

Perform the following actions to identify and correct the
failure:

After an Initiate Bare-metal discovery completes; HP SIM
acknowledges the server but it cannot be used for
installation, capture, or deployment operations.

Examine the task results with the HP SIM Tasks &
Logs

→View Task Results task to determine why the task

failed.

Start the remote console of the server you are trying to
discover. For information on the remote console,

“Connecting to a remote console” (page 187)

.

The sequence of events should be:

1. (For PXE boot only) A DHCP request. You should see

the target server issuing a DHCP request and
receiving a response.

If the target server does not receive a response, make
sure that the log file on the DHCP server contains a
DHCPDISCOVER

entry with the target server’s MAC

address.

25.15 Troubleshooting OS installation and image deployment problems 233

Advertising