1-
General
|
|
|
|
|
| 1-1.
Q: |
What is DRD?
|
| A:
|
Dynamic Root Disk (DRD) is an HP-UX system
administration toolset that
lets you create and modify an inactive system image without shutting
down
the system. To do this, you can clone the active system image and
modify the
cloned (inactive) system image while the system is running. When ready,
you
can boot the cloned image. Usually, the only downtime required is the
rebooting
process.
System administrators use DRD to manage
system images on HP PA-RISC
and Itanium®
-based systems.
DRD complements other parts of your total HP
solution by reducing system
downtime required to install and update patches and other software.
top
|
| 1-2.
Q: |
What HP-UX releases will DRD
run on?
|
| A:
|
This release of DRD runs on both Integrity
and PA platforms running either of the following operating systems:
- HP-UX 11i v2 (11.23) September 2004 or
more recent
- HP-UX 11i v3 (11.31)
top
|
| 1-3.
Q: |
How can I modify the inactive
system image without
affecting the active system image?
|
| A:
|
By only using DRD commands to modify the
inactive system image. DRD
commands are specially designed to create the inactive system image and
modify
only that image.
top
|
| 1-4.
Q: |
How can I be sure that the
clone is a consistent
system if the original system is still active?
|
| A:
|
Choose a time to create the system clone
when the booted system is fairly
quiet. Note that only the root volume group is cloned, so application
activity
on other volume groups will not affect the integrity of the clone. In
this
release of DRD, the copy operation is done by fbackup and frecover.
top
|
| 1-5.
Q: |
What are the DRD commands?
|
| A:
|
The seven DRD commands are:
| drd
clone |
| drd
runcmd |
| drd
activate |
| drd
deactivate |
| drd
mount |
| drd umount |
| drd status |
In addition, the drd runcmd command
allows you to
run specific Software Distributor (SD) commands on the inactive system
image
only. These SD commands are:
| swinstall |
| swremove |
| swlist |
| swmodify |
| swverify |
| swjob |
|
Two other commands can be executed by the drd runcmd command.
- The view
command can be used to view logs produced by commands that were
executed by drd runcmd.
- The kctune
command can be used to modify kernel parameters.
top
|
| 1-6.
Q: |
Are instance numbers changed
when the clone is
booted?
|
| A:
|
The ioconfig file and the entire /dev directory
are copied by the DRD clone operation, so instance numbers will not change when the clone is booted.
(On a system with an LVM root, the LVM information is modified so that the booted volume group is always vg00.
The /dev/vg00 directory is removed from the clone and the
/dev/drd00 directory is renamed /dev/vg00.)
top
|
| 1-7.
Q: |
Can the clone be mirrored?
|
| A:
|
Use the drd clone -x mirror_disk option to mirror the clone when you
create it. Also see the DRD Information Library Web page
for a white paper that explains how you can mirror the clone after the clone has been created.
top
|
| 1-8.
Q: |
I have many disks I would like
to clone in one
operation. Is the target a single disk?
|
| A:
|
Yes, for this release of DRD, the target
disk must be a single
disk.
top
|
| 1-9.
Q: |
What if the DRD contains more
than one disk? Does
DRD handle this?
|
| A:
|
Currently, the target disk must be a single physical disk, or SAN LUN, large enough to hold all of the root
volume file systems. This allows a customer to clone the root volume group even if it is spread across
multiple disks. Note that this is a one-way, many-to-one operation.
top
|
| 1-10.
Q: |
Can DRD clone all the
partitions; s1, s2, &
s3?
|
| A:
|
All partitions are created and s1 and s2 are
copied. This release
of DRD does not copy the HP service partition.
top
|
| 1-11.
Q: |
Does DRD work with both LVM
and VxVM root disks?
|
| A:
|
Yes, the root group being cloned can be managed by any release of LVM on an OS release supported by DRD.
In addition, the root group can be managed by VxVM 4.1 (HP-UX 11i v2 or 11i v3) or VxVM 5.0 (HP-UX
11i v2 only). See the DRD Downloads and Updates
Web page for information about the required patches if you are cloning a VXVM root. VxVM 5.0 on
HP-UX 11i v3 is not yet available. Please re-check the DRD
Downloads and Updates Web page periodically for updates.
If you are cloning a supported root, you can
have non-root groups on the system managed by any release of LVM or VxVM,
including VxVM 5.0. These groups are, of course, not cloned.
top
|
| 1-12.
Q: |
Does DRD support vPars?
|
| A:
|
Yes, DRD can clone the root volume of a
vPar, and the clone can be patched
and booted.
top
|
| 1-13.
Q: |
If swconfig is not
supported
by DRD, and swinstall
runs swconfig,
will it work properly?
|
| A:
|
DRD defers the configuration part of an
install operation, which remains
inactive until the system image is booted. (This behavior is similar to
what
happens when kernel software is installed.)
top
|
| 1-14.
Q: |
Does the cloned root disk (VG)
contain same VGID
as the original root VG?
|
| A:
|
The volume group will be vg00 when the clone
is booted. It will have
a different minor number than the original vg00.
top
|
| 1-15.
Q: |
How is lvmtab
adjusted to
reflect the new device name for the disk in vg00?
|
| A:
|
The copy of lvmtab
on the cloned image is modified
by the clone operation to contain information that will reflect the
desired
volume groups when the clone is booted.
top
|
| 1-16.
Q: |
Do the DRD administration
tools keep track
of the active boot disk's host attachment, in particular SAN boot
disks?
|
| A:
|
The system administrator is responsible for
choosing a disk that DRD
can use. This would include any communication with the SAN to assign
the disk
to a particular system.
top
|
| 1-17.
Q: |
What is the difference between
DRD Hot Recovery
and Mirror Disk/UX?
|
| A:
|
Mirror Disk/UX keeps an "always up to date"
image of the booted system.
Conversely, DRD provides a "point in time" image. The booted system and
the
clone may then diverge due to changes to either one. Keeping the clone
unchanged
is the Hot Recovery scenario. Note that DRD is not available for HP-UX
11.11,
which limits your options on those systems.
top
|
| 1-18.
Q: |
Since I take a regular root
disk and use it to
boot another system, can I use the cloned disk I created with DRD to
boot
another system?
|
| A:
|
It is possible to do this, however, factors
such as machine personality
(e.g., hostname, IP address, etc.) make this very difficult. HP does
not recommend
using the cloned disk to boot another system.
top
|
| 1-19.
Q: |
What happens during a clone
copy if, for example, /opt/xxxxx
is
mounted from /dev/vg01/lvolxxxx?
Is the content of /opt/xxxxx
copied
to the clone or does it only create its directory?
|
| A:
|
Only the contents of vg00 are copied. In the
above example, the content
of /opt/xxxxx will not be
copied to the clone. A system
that has the /opt file
systems (or any file system that
is patched) not in vg00 is not suitable for use with DRD.
top
|
| 1-20.
Q: |
What are the effects of the
clone copy on system
operation (for example, performance)?
|
| A:
|
The I/O of the clone operation will have
some impact, particularly if
the disk is close to the booted system. For example, if the two disks
are
on the same SCSI chain, then I/O to the clone will affect performance
of I/O
on the booted system. DRD's performance is similar to system
performance when
using Ignite to create recovery images, which many system
administrators find
acceptable.
top
|
| 1-21.
Q: |
IgniteUX offers net and tape
images. What is the
advantage of DRD?
|
| A:
|
DRD provides a
“make-disk-recovery” solution and has several
advantages over IgniteUX net and tape images:
-
-
No impact on
network performance will occur
-
No security
issues of transferring data across the network
will occur
top
|
| 1-22.
Q: |
How can I be sure the patches
I install are safe in a DRD environment?
|
| A:
|
You can install patches packaged in standard
Software Distributor (SD) format. Most patches for HP–UX
11.23 September, 2004 and above are DRD safe. However, the DRD product
maintains a list of those patches known to be unsafe. When the revision
of SD required by DRD is run by drd
runcmd, it makes sure that a patch does not
appear in the list before installing or removing it.
For more information about DRD-unsafe
patches, see the Managing Rare DRD-Unsafe Patches White Paper
at the Dynamic
Root Disk documentation page.
top
|
2-
Using Dynamic Root Disk Commands
|
|
|
| 2-1.
Q: |
How do I use DRD commands?
|
| A:
|
In some cases, you simply enter the command.
For example, if you enter drd
mount, the command mounts the inactive system
image file set. In
other cases, you need to use one or more command options. For example, drd
clone always needs the -t
option with the target
disk identity.
You can also enter the drd runcmd followed
by an
SD command. For example,
drd
runcmd swlist.
drd
runcmd limits the action of swlist to
just the inactive system image.
For more information about these commands,
see the Dynamic
Root Disk Administration Guide for HP-UX 11i (B11.23) on the Information Library
page.
top
|
| 2-2.
Q: |
Will I need to reboot while
installing DRD?
|
| A:
|
On HP-UX 11i v3 (11.31) systems, there are
no required patches for DRD so installing DRD will not cause a reboot.
Neither DRD nor its corequisites require a
reboot. However, one of DRD's corequisites, patch PHCO_34195, does have
a further corequisite on two other patches, PHKL_33024 and PHKL_33025,
that do require a reboot. Most systems will already have these two
patches installed and will have no need for a reboot.
To determine definitely whether your
installation of DRD will require
a reboot, preview the installation and check whether any kernel patches
(PHKL_*)
are included in the selection at the end of the analysis phase of the
install.
top
|
| 2-3.
Q: |
How can I tell if I have
chosen a sufficiently large target disk for a DRD clone?
|
| A:
|
A simple mechanism for determining if you
have chosen a sufficiently large disk is to run a preview of the drd clone command:
drd
clone -p -v -t path_to_block_DSF
Where path_to_block_DSF
is of the form:
- HP-UX 11i v2: /dev/disk/cXtXdX
- HP-UX 11i v3: /dev/disk/diskX
The preview operation includes the disk
space analysis needed to see
if the target disk is sufficiently large.
top
|
| 2-4.
Q: |
If I am running multiple drd runcmd
operations,
can I avoid mounting and unmounting the inactive system image for each runcmd execution?
|
| A:
|
You can avoid multiple mounts and unmounts
by using drd mount
to
mount the inactive system image before the first runcmd operation
and drd umount
to unmount the inactive system image after
the last runcmd
operation. When drd runcmd
finds
the file systems in the clone already mounted, it does not unmount them
(nor
will it export the volume group) at the completion of the runcmd operation.
top
|
| 2-5.
Q: |
Does the DRD clone operation
verify that the target
is not in use or part of another VG? Once the root disk is cloned, is
there
anything in the operating system that prevents the non-active target
from
being reused? How can inactive targets be identified?
|
| A:
|
If the disk is currently in use by another
volume group that is visible
on the system, the disk will not be used. If, on the other hand, the
disk
contains LVM, VxVM, or boot records but is not in
use
on the system, you must use the -x
overwrite option to
tell DRD to overwrite the disk. Already-created clones will contain
boot records;
a preview using the drd
mount or drd
activate commands
will show the disk that is currently in use as an inactive system image.
top
|
| 2-6.
Q: |
Can I vgimport the target
disk
as /dev/vgtmp?
|
| A:
|
The drd
mount command will vgimport the
target disk, vgchange
it to make it active, and mount all
the file systems in the cloned volume group. However, these operations
do
not use the lvmtab on the
clone itself. After booting
the clone, you can also mount the original system image. (You can also vgimport the
clone manually, but it is much easier to use the drd mount command.).
top
|
| 2-7.
Q: |
Can DRD show whether DRD is configured or not configured?
|
| A:
|
You can run the drd status command to determine whether a clone has been
created. The drd status command output shows, among other data, the
disk that was cloned and the target disk of the drd clone operation.
For further information on the drd status command, see the
drd_status(1M) manpage. To determine the validity of all software installed on the inactive
system image, execute:
drd runcmd swverify \*
top
|
| 2-8.
Q: |
Some processes in DRD,
particularly drd
clone and drd
runcmd, take a significant amount
of time. Can these processes be interrupted?
|
| A:
|
All DRD processes, including drd clone and drd
runcmd, can be safely interrupted issuing
Control/C (SIGINT) from
the controlling terminal or by issuing kill -HUP<pid>
(SIGHUP).
This action causes DRD to abort processing and perform any necessary
clean
up. Do not interrupt DRD using the kill
-9 <pid>
command
(SIGKILL), which fails to abort safely and does not perform cleanup.
top
|
3-
Troubleshooting Dynamic Root Disk
|
|
|
| 3-1.
Q: |
Why can't I see the inactive
system image file
systems when I run the bdf
command?
|
| A:
|
You have to mount the inactive system image
before you can see the file
systems using bdf.
Use the drd mount
command
to mount the inactive system image.
top
|
| 3-2.
Q: |
When I run DRD, what
information should I collect
before calling HP about a problem?
|
| A:
|
Before you contact HP, be sure to gather the
following useful
information:
-
Find the
HP-UX revision of your system by entering:
uname -a
-
Look at the
contents of the log files under /var/opt/drd.
It may be helpful to have the data for the entire session. In DRD log
files,
sessions are delineated with ========. If there is too much data in the
log,
locate the first sign of trouble and the next several error messages.)
-
Make
available information from the swagent.log
files
stored in depots (for example, /var/spool/sw/swagent.log).
The swagent daemon
creates these when it
reads or writes from a depot containing information written by swagent.
Depots at other locations have a similarswagent.log.
-
If you are
troubleshooting a problem with drd
runcmd,
you may also want to consult the contents of the logfiles on the
inactive
system image located at:
/var/opt/drd/mnts/sysimage_001/var/adm/sw/sw*.log
or
/var/opt/drd/mnts/sysimage_000/var/adm/sw/sw*.log
top
|
| 3-3.
Q: |
I can't remember which disk was used as the target of my clone operation.
How can I identify it?
|
| A:
|
The disk that was used for the clone is displayed by the
drd status command.
drd status
The Clone Disk field specifies the target disk used to clone the original image.
top
|
| 3-4.
Q: |
I issued a drd activate command
but changed my mind and do not want to boot the clone. How can I undo
the drd
activate command?
|
| A:
|
If you run drd
activate and then decide not to boot the
inactive image on the next reboot, use the drd deactivate
command to set the primary boot disk to the currently booted disk.
For further information on the drd deactivate
command, see the drd-deactivate(1M) manpage.
top
|
| 3-5.
Q: |
Can I change file system sizes
when I create a
clone?
|
| A:
|
This release of DRD does not provide a
mechanism for resizing
file systems during a drd
clone operation. However, after
the clone is created, you can manually change file system sizes on the
inactive
system without needing an immediate reboot. The whitepaper, Using
the Dynamic Root Disk Toolset describes resizing file systems
other than /stand. The
whitepaper Using the
DRD toolset to extend the /stand file system in an LVM environment
describes
resizing the boot (/stand)
file system on an inactive
system image. Both of these documents can be found on the Information Library.
top
|
| 3-6.
Q: |
How will fbackup|frecover
handle “busy”
files?
|
| A:
|
DRD does utilize fbackup and frecover, however, we
have not encountered a problem with busy files. While frecover can fail to
overwrite busy files, this is not a problem in DRD because it is
cloning a fresh volume group. fbackup
can fail if files are changing during the fbackup operation,
but this is more common with spool files and log files. If you are
concerned, use the drd
runcmd swverify \* command to check the
integrity of the clone after it is created.
top
|
| 3-7.
Q: |
What happens with the
configuration scripts if
I install patch A and later install patch B, which then supersedes
patch A?
|
| A:
|
If you install both Patch A and Patch B
before booting, only Patch B
is configured. This should not present a problem unless you later
remove patch
B with swremove,
thereby exposing Patch A in an unconfigured
state. You would then need to swconfig
or reinstall Patch
A.
top
|
| 3-8.
Q: |
If I use the IgniteUX server;
is it aware of the
DRD clone?
|
| A:
|
The Ignite server will only be aware of the
clone if it is mounted during
a make_*_recovery
operation. Most customers will probably
prefer to keep the clone unmounted so that only the active system image
is
archived by Ignite. In addition, the customer will probably want to
keep a
record of the clone hardware path to enable Hot Recovery using DRD.
top
|
Printable version
