| ETERNUS SF AdvancedCopy Manager Operator's Guide 13.0 -AIX- |
|
Contents
Index
|
This chapter provides notes on the operations of AdvancedCopy Manager.
This chapter provides notes on the backup and replication operation.
This chapter provides general notes on the backup and replication operation.
The following devices must not be subject to the backup or replication operation.
Device on which the system is stored
Device on which AdvancedCopy Manager has been installed
Device on which the resource files of AdvancedCopy Manager exists
|
Data on transaction volumes |
Data consistency |
Action to take |
|---|---|---|
|
File system |
The commands of AdvancedCopy Manager unmount the filesystem and the correspondence is secured. |
Please refer to 'Preprocessing and Postprocessing of Backup and Restoration' and 'Preprocessing and Postprocessing of Replication' of this manual. |
|
Other than above |
It is necessary to secure the correspondence of data by operation. |
Please deal with the business of the stop etc. when you execute the replication the backup. |
Before starting the backup or replication operation, fetch device information managed by all the storage servers on the Web window of AdvancedCopy Manager. This operation will take time proportional to the total number of devices defined on the selected storage servers. If many devices are involved, perform this operation while the CPU load or I/O load is low.
As a guide, about 0.5 seconds are required per device (disk) under a no-load condition.
When the device is being made multipass composition, even if one of the two's path is blockaded, a path shift is not carried out automatically.
When one of the two's path is blockaded, please re-perform backup or replication after the following management implementation.
The following commands are carried out and a path is switched. Please execute the following commands to a normal device.
/usr/sbin/lspv hdisk*
A backup or replication command is executed again.
Before backup/restoration or replication of a volume that can be mounted from multiple servers, cancel mounting from the other servers.
For disks that need not be mounted from other servers, prevent multiple servers from detecting or accessing the same logical disk by setting up hardware, such as devices in the ETERNUS storage system and Fibre Channel switches, accordingly.
For a copy operation targeted at a file system, unmount the applicable volume to prevent data access and maintain the data integrity.
If backup/restoration and replication are performed without unmounting the volume, an error occurs since a volume in use cannot be unmounted.
Note the following about the processing in such a state in which unmounting is enabled:
Another application is not using the volume. If another application is using it, suspend the application temporarily.
Users are not using the volume. If users are using it, prohibit them from using it temporarily.
Another volume is not mounted on the volume. If it is mounted, unmount it temporarily.
The volume is not shared with NFS sharing (for Solaris/HP-UX/Linux). If it is shared, unshare it temporarily.
Unmounting is required only during command execution. Operation can be resumed after command execution.
This chapter provides notes on the snapshot high-speed backup.
The snapshot high-speed backup requires as many backup volumes as the number of specified generations. Thus, the snapshot backup is disabled if:
All the specified number of generations have been backed up, and
No new volume that can be used as a backup volume exists.
Performing the snapshot high-speed backup continuously with the same transaction volume specified will perform the backup operation in parallel.
Performing the snapshot high-speed backup continuously more often than the number of preservation generations will cancel the backup operations starting from the oldest one. In other words, you cannot simultaneously perform the backup operation more often than the specified number of generations.
This chapter includes some notes on synchronous high-speed backup.
The backup execution command cannot be executed before the transaction and backup volumes have become equivalency holding state.
Backup policies can be set if the number of registered backup volumes is enough for snapshot-type high-speed backup operation even if the number is not enough for synchronous high-speed backup operation. In this event, synchronized backup may not be possible.
This chapter provides notes on restoration.
If data on a transaction volume has been rewritten since the latest backup data is collected to a backup volume and until restoration is performed, the rewritten data shall not be guaranteed.
AIX uses LVM to manage all devices. Therefore, executing backup rewrites the LVM management information in the backup volume with that in the transaction volume. As a result, the backup volume cannot be activated and the logical volume of the backup volume cannot be mounted.
To mount the backup volume, the LVM management information in the backup volume must be rewritten. However, this operation is not recommended because it results in an inability to use the normal procedure to restore the transaction volume. If circumstances require mounting of the backup volume, the procedure given below can be used to mount the backup volume. Follow the procedure very carefully; otherwise, transaction data may be destroyed. For information on each command, refer to the AIX manuals. Also, see "Restoration procedure after LVM management information in the backup volume is rewritten" in the next part of this section.
Using, for example, the lspv command, specify the physical volume name of the backup volume.
# /usr/sbin/lspv hdisk0 0004f10aa92e686c rootvg hdisk1 0004f10a1c7879c5 vg01 |
In the above example, the volume group name of the backup volume is vg01, and the physical volume name is hdisk1.
Use the chdev command to temporarily remove the backup volume from LVM.
# /usr/sbin/chdev -l hdisk1 -a pv=clear |
Use the exportvg command to export the backup volume.
# /usr/sbin/exportvg vg01 |
Use the recreatevg command to rewrite the LVM management information in the backup volume.
# /usr/sbin/recreatevg -y vg01 hdisk1 |
Executing the recreatevg command changes the logical volume name in the volume group. Use the lsvg command to confirm the new logical volume name.
# /usr/sbin/lsvg -l vg01 vg01: LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT fslv01 jfs 2 2 1 closed/syncd N/A |
Mount a file system by specifying the logical volume name confirmed in step 5. Unmount the logical volume before the next time that backup is performed. If the logical volume in the backup volume is mounted, backup fails.
Rewriting the LVM management information in the backup volume makes it inconsistent with that in the transaction volume and prevents use of the normal procedure to restore the backup volume. In such cases, follow the procedure given below to restore it. Follow the procedure very carefully; otherwise, transaction data may be destroyed. For information on each command, refer to the AIX manuals.
Refer to "Preprocessing and Postprocessing of Backup and Restoration," and edit the post-processing script for restoration.
Execute restore processing. The transaction volume is now inactive because of the action taken in step 1.
Using, for example, the lspv command, specify the physical volume name of the transaction volume.
# /usr/sbin/lspv hdisk0 0004f10aa92e686c rootvg hdisk1 0004f10a1c7879c5 vg01 |
In the above example, the volume group name of the transaction volume is vg01, and the physical volume name is hdisk1.
Use the chdev command to temporarily remove the transaction volume from LVM.
# /usr/sbin/chdev -l hdisk1 -a pv=clear |
Use the exportvg command to export the transaction volume.
# /usr/sbin/exportvg vg01 |
Use the recreatevg command to rewrite the LVM management information in the transaction volume.
# /usr/sbin/recreatevg -y vg01 hdisk1 |
Executing the recreatevg command changes the logical volume name in the volume group. Use the lsvg command to confirm the new logical volume name.
# /usr/sbin/lsvg -l vg01 vg01: LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT fslv01 jfs 2 2 1 closed/syncd N/A |
Mount a file system by specifying the logical volume name confirmed in step 7.
When a backup volume is restored to a volume other than a transaction volume, the pre-processing and post-processing scripts for restoration are not executed. This operation also makes the LVM management information in the backup volume inconsistent with that in the restoration destination volume. Accordingly, the LVM management information must be rewritten. For these reasons, to restore a backup volume to a volume other than a transaction volume, follow the procedure given below. Follow the procedure very carefully; otherwise, transaction data may be destroyed. For information on each command, refer to the AIX manuals.
Unmount all logical volumes in the volume group of the restoration destination volume.
Use the varyoffvg command to deactivate the volume group of the restoration destination volume. The execution example shown below uses volume group named vg01 for the restoration destination volume.
# /usr/sbin/varyoffvg vg01 |
Execute restoration for the restoration destination volume. For the procedure, see "Performing restoration."
Using, for example, the lspv command, note the physical volume name of the restoration destination volume.
# /usr/sbin/lspv hdisk0 0004f10aa92e686c rootvg hdisk1 0004f10a1c7879c5 vg01 |
In the above example, the volume group name of the restoration destination volume is vg01, and the physical volume name is hdisk1.
Use the chdev command to temporarily remove the restoration destination volume from LVM.
# /usr/sbin/chdev -l hdisk1 -a pv=clear |
Use the exportvg command to export the restoration destination volume.
# /usr/sbin/exportvg vg01 |
Use the recreatevg command to rewrite the LVM management information in the restoration destination volume.
# /usr/sbin/recreatevg -y vg01 hdisk1 |
Executing the recreatevg command changes the logical volume name in the volume group. Use the lsvg command to confirm the new logical volume name.
# /usr/sbin/lsvg -l vg01 vg01: LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT fslv01 jfs 2 2 1 closed/syncd N/A |
Mount a file system by specifying the logical volume name confirmed in step 8.
Note the following general points on cluster operation:
Do not make a SWSTGNODE registration that configures an environment variable to the entire system.
AdvancedCopy Manager does not monitor daemons composing a transaction. If the daemon of a transaction stops for some reason, AdvancedCopy Manager does not automatically restart it, so you must restart it manually. Moreover, the stopping of a transaction-oriented daemon of AdvancedCopy Manager does not cause the package to fail over.
For details on the daemons composing AdvancedCopy Manager transactions and how to start these daemons, see "Starting and Stopping Daemons" in this manual.
For backup operation in a cluster system, note the following points that do not apply to regular operation:
If failover occurs while running a backup command, you must secure consistency by using the resource adjustment command as a recovery measure.
To use synchronous high-speed backup, do not register a backup to any cluster transaction.
You must perform backup operation on the active system of the service group because AdvancedCopy Manager is placed into the service group and operates as part of the service group.
You cannot make a backup from a standby node or another service group.
To perform backup/restoration by using AdvancedCopy Manager, you usually must unmount transaction volumes beforehand and mount the volumes later.
If the mount point of transaction volume have been registered to a cluster transaction, set the mount point resource online or offline from the cluster transaction instead of mounting or unmounting the volumes. Alternatively, modify the backup/restoration preprocessing and postprocessing scripts so that the transaction volumes are set offline or online instead of being unmounted or mounted.
Moreover, the Volume is actually after doing offline/online of mount point resource and there will be a time difference by the time mount/unmount is done. Therefore, please add the processing (The Result of sleep and the df Command is observed) meets until mount/unmount is actually done after the location where the success or failure of offline/online is judged.
For details and examples on the preprocessing and postprocessing scripts, see "Backup/Restoration Preprocessing/Postprocessing" in this manual. For details on the commands usable to set resources offline and online, see the relevant cluster system manual.
When a service group to which AdvancedCopy Manager belongs is operating, backup operation is enabled only on the active node. The standby node environment is not ready (to provide necessary resources) for backup operation, thus disabling backup operation. Likewise, if the service group has been stopped, the environment is not ready and disables backup operation.
However, only when the service group has been stopped is backup operation enabled by temporarily arranging the minimum essentials.
If any of the following types of shared disks cannot be enabled (i.e., set online or mounted in a system), backup operation is disabled:
Shared disk for AdvancedCopy Manager shared data
Backup operation disk (transaction volume/backup volume to operate)
The following operations are not allowed in backup operation:
Backup/restoration through GUI
Manipulation for backup operation from a storage management server (host specification with the -h option)
Addition or deletion of a transaction volume/backup volume, or change in policy
To execute a display command for backup management to a storage management server transaction that also functions as a storage server transaction, you must specify options for the command. For details on the display command for backup management, see "Commands for Backup Management" in the ETERNUS SF AdvancedCopy Manager Operator's Guide.
If the service group has been stopped, perform backup operation by taking the following steps:
Confirm that the service group has been stopped on both nodes.
For details on how to stop a service group, see the VERITAS Cluster Server manual.
Log in to the node on which you want to make a backup through telnet, for example.
You cannot use the logical IP address. Directly use a node that is operable by using the physical IP address.
Enable the shared disk.
Start (set online) the shared disk.
Mount the shared disk for AdvancedCopy Manager shared data.
When the transaction volume is a file system, mount it.
Be sure to enable the shared disk from either node. Do not enable the shared disk from both nodes.
Perform backup operation.
You can perform backup operation in the same way as for regular operations for cluster operation.
You can use the restore command to perform restoration.
Release all the shared disks you enabled in Step 3.
Unmount the file systems you mounted.
Stop (set offline) the shared disk.
Start (set online) the service group.
Start the service group, if necessary.
For details on how to start a service group, see the VERITAS Cluster Server manual.
For replication operation in a cluster system, note the following points that do not apply to regular operation:
If failover occurs while running a replication command, you must secure consistency by using the resource adjustment command as a recovery measure.
When you use synchronous replication, do not register any destination volume to a cluster transaction.
You must perform replication operation on the active system of the service groups because AdvancedCopy Manager is placed in the service groups and operates as part of the service groups.
You cannot perform replication from a standby node or another service group.
To perform replication, you must unmount original/source volumes beforehand and mount the volumes later for some commands.
If the mount points of original/source volumes have been registered to a cluster transaction, set the mouont point resources original/source volumes online or offline from the cluster transaction instead of mounting or unmounting the volumes. Alternatively, modify the unmount and mount processes in the replication preprocessing and postprocessing scripts to offline and online processes, respectively.
Moreover, the Volume is actually after doing offline/online of mount point resource and there will be a time difference by the time mount/unmount is done. Therefore, please add the processing (The Result of sleep and the df Command is observed) meets until mount/unmount is actually done after the location where the success or failure of offline/online is judged.
For details and examples on the types of commands that require the volumes to be unmounted and mounted, and the preprocessing and postprocessing scripts, see "Preprocessing and Postprocessing of Replication" in this manual. For details on the commands usable to set resources online and offline, see the relevant cluster system manual.
When a service group to which AdvancedCopy Manager belongs is operating, replication operation is enabled only on the active node. The standby node environment is not ready (to provide necessary resources) for replication operation, thus disabling replication operation. Likewise, if the service group has been stopped, the environment is not ready and disables replication operation.
However, only when the service group has been stopped is replication operation enabled by temporarily arranging the minimum essentials.
If any of the following types of shared disks cannot be enabled (i.e., set online or mounted in a system), replication operation is disabled.
Shared disk for AdvancedCopy Manager shared data
Replication operation disks (source volume/destination volume to operate)
The following operations are not allowed in replication operation:
Manipulation for replication operation from a storage management server (host specification with the -h option)
Addition or deletion of an original/replica volume
Inter-server replication without the -m option specified
To execute a display command for replication management to a storage management server transaction that also functions as a storage server transaction, you must specify options for the command. For details on the display command for replication management, see "Commands for Replication Management" in this manual.
Perform replication operation during stoppage of the service group by taking the steps below.
Confirm that the service group has been stopped on both nodes.
For details on how to stop a service group, see the VERITAS Cluster Server manual.
Log in to the node on which you want to perform replication operation through telnet, for example.
You cannot use the logical IP address. Directly use a node that is operable by using the physical IP address.
Enable the shared disk.
Start (set online) the shared disk.
Mount the shared disk for AdvancedCopy Manager shared data.
When the original volume and replica volume have been mounted as file systems for operation, mount the volumes.
Be sure to enable the shared disk from either node. Do not enable the shared disk from both nodes.
Perform replication operation.
You can perform replication operation in the same way as for regular operations for cluster operation.
You can also perform restoration from the replica volume to the original volume.
Release all the shared disks you enabled in Step 3.
Unmount the file systems you mounted.
Stop (set offline) the shared disk.
Start (set online) the service group.
Start the service group, if necessary.
For details on how to start a service group, see the VERITAS Cluster Server manual.
For notes on the Web screen environment, see "Setting up a client" in the ETERNUS SF AdvancedCopy Manager User's Guide.
|
Contents
Index
|