01-10 Backing Up and Restoring the M2000

10 Backing Up and Restoring the M2000

About This Chapter

This task is performed to back up and restore M2000 data.

10.1 M2000 Backup and Restore SolutionsThis part describes the M2000 backup and restore solutions. Based on the types of backup data,the M2000 provides three backup solutions: dynamic data backup, static data backup, and systemdata backup. Dynamic data backup is applicable to the dynamic service data of the M2000system. Static data backup is applicable to the data of the M2000 server applications, the dataof the database applications, the data of the database files and the data of M2000 configurationfiles. System data backup is applicable to the system data on the M2000 server. The system dataon the M2000 server includes the data of the operating system, the data of database applications,and the data of M2000 server applications.

10.2 Backing Up and Restoring M2000 Dynamic DataThis task is performed to back up and restore M2000 dynamic data.

10.3 Backing Up and Restoring M2000 Static DataThis task is performed to back up and restore the M2000 static data.

10.4 Backing Up and Restoring M2000 System DataThis task is performed to back up and restore the M2000 system data. The M2000 provides twobackup and restore solutions for backing up and restoring the system: ufsdump and Veritas. Theufsdump solution backs up the data on the system disk to the tape drive using the Solaris systemcommands and scripts. The ufsdump solution is applicable to all server types. The Veritassolution backs up data through the backup and restore components (NetBackup and BMR)provided by the Symantec company. The Veritas solution is applicable to all server types. Fordetails, refer to the manuals related to the Veritas system backup solution.

M2000Administrator Guide (S10) 10 Backing Up and Restoring the M2000

Issue 04 (2007-06-08) Huawei Technologies Proprietary 10-1

10.1 M2000 Backup and Restore SolutionsThis part describes the M2000 backup and restore solutions. Based on the types of backup data,the M2000 provides three backup solutions: dynamic data backup, static data backup, and systemdata backup. Dynamic data backup is applicable to the dynamic service data of the M2000system. Static data backup is applicable to the data of the M2000 server applications, the dataof the database applications, the data of the database files and the data of M2000 configurationfiles. System data backup is applicable to the system data on the M2000 server. The system dataon the M2000 server includes the data of the operating system, the data of database applications,and the data of M2000 server applications.

10.1.1 Backup of M2000 Dynamic DataThis part describes the backup of M2000 dynamic data.

10.1.2 Backup of M2000 Static DataThis part describes the backup of M2000 static data.

10.1.3 Backup of M2000 System DataThis part describes the backup of the M2000 system data.

10.1.4 Policies of M2000 Data BackupThis part describes the policies of M2000 data backup. The policies include the scenarios fordata backup, and the rules for naming backup files, and the rules for naming the tapes that storethe backup data.

10.1.5 Scenarios of M2000 Data RestorationThis part describes the scenarios of the M2000 data restoration.


The dynamic data to be backed up refers to the dynamic service data of the M2000 system. Whenthe existing M2000 system works properly, you can use the dynamic backup data to restore thehistory data. Table 10-1 lists the content, mode, and storage device of the backup of dynamicdata.

Table 10-1 Dynamic data backup

Item Description

Content l Data in the databases: omcdb, fmdb, pmdb, sysmdb, itfndb, inrptdb2,omclogdb, omcsmdb, and omctmdb.

l M2000 system files: files in the paths /export/home/omc/var and /export/home/sysm.

10 Backing Up and Restoring the M2000M2000

Administrator Guide (S10)

10-2 Huawei Technologies Proprietary Issue 04 (2007-06-08)

Item Description

Backup mode Periodicbackup

After periodic backup is activated, periodic backup isperformed every day. On the specified day of each week,the system backs up all the involved data (full backup).At the same time of other six days, the system backs upthe new data (incremental backup). After a new periodicbackup is performed, the existing backup files areautomatically deleted.

Manualbackup

l Full backup: backs up all the dynamic data. After fullbackup is performed, all the backup files in the backupdirectory are deleted automatically. Full backup isperformed as required.

l Incremental backup: backs up only the changeddynamic data after the previous full backup. Thebackup files of the incremental backup, which aresaved in another folder, do not overwrite the backupfiles of the full backup. The incremental backup isperformed as required.

Storage device l Tapes

l Disks

l Both tapes and disks

Path of backupfiles

/export/home/backup/omc.

Amount ofrequired tapes

Dynamic data is backed up periodically. Each time before dynamic datais backed up, you need to prepare at least a 72 GB tape. You can also reusethe tapes if the backup files on them are no longer required

After you perform a backup task, the system automatically packs the backup content into abackup file.


The static data to be backed up involves the data in /opt/OMC, /opt/sybase, and /export/home ofthe hard disk on the M2000 server. The backup data is used for restoring the database or theM2000 program. Table 10-2 lists the content, partition, mode, and storage device of the backupof static data.



Table 10-2 Static data backup

Item Description

Content l M2000 server applications

l Database software

l Database files

l M2000 configuration files

Partition l /opt/OMC

l /opt/sybase

l /export/home

Backup mode One-time backup mode. Generally, the dataneed not be backed up again after it is backedup for the first time. You, however, mustback up the data again if you upgrade adatabase program, expand the capability ofthe database, or upgrade M2000 applicationprograms.

Storage device Tapes

Amount of required tapes In general, two 72 GB tapes are required forthe backup of static data. In practice,determine the amount of required tapesaccording to the used space of the files in aspecific partition or a path.NOTE

l You can use the command du -sh /opt/OMCto view the used space of the files in the /opt/OMC path.

l You can use the command du -sh /opt/sybase to view the used space of the files inthe /opt/sybase path.

l You can use the command df -h /export/home to view the used space of the files inthe /export/home partition.


The M2000 system data is backed up for restoring the existing operating system, databasesoftware, and M2O00 server applications. Table 10-3 lists the content, partition, mode, andstorage device of the backup of system data.




Table 10-3 System data backup

Item Description

Content l Solaris

l Database software

l M2000 server application software

Partition /

Backup mode One-time backup. The system data is backed up once forall unless a new operating system patch is installed.


Amount of required tapes In general, a 72 GB tape is required for the backup ofsystem data. In practice, determine the amount of requiredtapes according to the used space of the files in a specificpartition.NOTE

You can use the command df -h / to view the used space of thefiles in the / partition.

10.1.4 Policies of M2000 Data BackupThis part describes the policies of M2000 data backup. The policies include the scenarios fordata backup, and the rules for naming backup files, and the rules for naming the tapes that storethe backup data.

Typical Scenarios for M2000 Data BackupTable 10-4 describes the typical scenarios for M2000 data backup.

Table 10-4 Typical scenarios for M2000 data backup

No. Scenario Data

1 The M2000 system is installedfor the first time.

l Static data (manual backup)

l Dynamic data

l System

2 The M2000 program isupgraded, or the database isupgraded or expanded.


l Dynamic data

3 The operating system isupgraded. For example, a patchof the operating system isinstalled.

System

4 Routine backups. Static data (manual backup)



Rules for Naming Backup Files of Dynamic Data

NWhen backing up the static data, name the backup files in the following format: /export/home/backup/omc/*-YYYYMMDDhhmmss.tar. Where, "*" indicates all for the full backup or inc forincremental backup.

For instance:

l The file name all-20040601042002.tar refers to a full backup file generated at 04:20:02 onJune 1 of 2004.

l The file name inc-20040602042001.tar refers to an incremental backup file generated at04:20:01 on June 2 of 2004.

NOTE

l The latest backup information is recorded in the /export/home/backup/omc/backup.log file.

l During the backup process, the inc-YYYYMMDDhhmmss or all-YYYYMMDDhhmmss folder isgenerated in the /export/home/backup/omc path for storing backup files. When the backup iscomplete, the folder is packed as the inc-YYYYMMDDhhmmss.tar file, and the inc-YYYYMMDDhhmmss or all-YYYYMMDDhhmmss folder is automatically deleted.

Rules for Naming the Tapes Storing the Backup Data

Name the tapes that store the backup data in the *Backup tape (a of b) format. The character *stands for the following data:


l Dynamic data

l System data

For example, a tape is named dynamic backup tape (1 of 2). This tape name means that thedynamic data is backed up to two tapes and this tape is the first one.

10.1.5 Scenarios of M2000 Data RestorationThis part describes the scenarios of the M2000 data restoration.

Table 10-5 describes the data restore solutions in typical scenarios.

Table 10-5 Data restore solutions in typical scenarios

No. Scenario Data Restore Solution

1 During operation, theM2000 system is requiredto revert to a previous state,such as the state in the lastweek.

Dynamic data restoration

2 The database files or theM2000 configuration filesare damaged and theSolaris of the M2000 serverworks properly.

Static data restoration (The procedurefor restoring the dynamic data is alsoincluded.)




No. Scenario Data Restore Solution

3 The M2000 server breaksdown, and the Solarisoperating system cannot bestarted.

System restoration (The procedures forrestoring the dynamic data and staticdata are also included.)

10.2 Backing Up and Restoring M2000 Dynamic DataThis task is performed to back up and restore M2000 dynamic data.


10.2.2 Setting the Storage Device of Backup DataThis task is performed to set the storage device of backup data.

10.2.3 Backing Up M2000 Dynamic Data PeriodicallyThis task is performed to periodically back up the M2000 dynamic data.

10.2.4 Backing Up M2000 Dynamic Data ManuallyThis task is performed to manually back up the M2000 dynamic data.

10.2.5 Restoring M2000 Dynamic DataThis task is performed to restore M2000 dynamic data.


The dynamic data to be backed up refers to the dynamic service data of the M2000 system. Whenthe existing M2000 system works properly, you can use the dynamic backup data to restore thehistory data. Table 10-6 lists the content, mode, and storage device of the backup of dynamicdata.

Table 10-6 Dynamic data backup

Item Description

Content l Data in the databases: omcdb, fmdb, pmdb, sysmdb, itfndb, inrptdb2,omclogdb, omcsmdb, and omctmdb.

l M2000 system files: files in the paths /export/home/omc/var and /export/home/sysm.



Item Description

Backup mode Periodicbackup

After periodic backup is activated, periodic backup isperformed every day. On the specified day of each week,the system backs up all the involved data (full backup).At the same time of other six days, the system backs upthe new data (incremental backup). After a new periodicbackup is performed, the existing backup files areautomatically deleted.

Manualbackup

l Full backup: backs up all the dynamic data. After fullbackup is performed, all the backup files in the backupdirectory are deleted automatically. Full backup isperformed as required.

l Incremental backup: backs up only the changeddynamic data after the previous full backup. Thebackup files of the incremental backup, which aresaved in another folder, do not overwrite the backupfiles of the full backup. The incremental backup isperformed as required.

Storage device l Tapes

l Disks

l Both tapes and disks

Path of backupfiles

/export/home/backup/omc.

Amount ofrequired tapes

Dynamic data is backed up periodically. Each time before dynamic datais backed up, you need to prepare at least a 72 GB tape. You can also reusethe tapes if the backup files on them are no longer required

After you perform a backup task, the system automatically packs the backup content into abackup file.

10.2.2 Setting the Storage Device of Backup DataThis task is performed to set the storage device of backup data.

PrerequisiteYou have logged in to the M2000 system as the root user.

ContextThe backup files for the dynamic data backup are stored on hard disks. To store the backup fileson a tape, modify the relevant settings in the cmserversvc.xml file.

During the setting of the storage device, the M2000 services are stopped, and the performanceand alarm data of the managed NEs are not processed.




CAUTIONl Avoid modifying the storage device file when the backup is in process.

l The backup overwrites the backup files that already exist on the disk.

l If you set the backupMedia parameter to tape or all, replace the tape after the backup iscomplete.

l Avoid replacing the storage device during the backup process.

l For the purpose of security, back up the dynamic data to the tape.

l When restoring the static data and system, use the dynamic data backup files on the tapeto restore the latest service data.

Procedure

Step 1 Switch to the M2000 installation path.The default M2000 installation directory is /opt/OMC.

# cd /opt/OMC

Step 2 Stop the M2000 services.

# . ./svc_profile.sh

# stop_svc

Step 3 Switch to the directory /export/home/omc/var/etc/conf, and then run the vi command to viewthe cmserversvc.xml file.

# cd /export/home/omc/var/etc/conf

# vi cmserversvc.xml

Step 4 Search for the line <param name="backupMedia">, and then change the disk for both the fulland incremental backups to the actual storage device.<module name="all">...<param name="backupMedia">disk</param>

and

<module name="inc">...<param name="backupMedia">disk</param>

Table 10-7 lists the values for the parameter backupMedia.

Table 10-7 Values for backupMedia

Storage Device Value

Hard disks disk

Tapes tape

Hard disks and tapes all



Step 5 To back up dynamic data on a tape, specify the path of the tape drive. To back up the dynamicdata on the disk, go to Step 6.

Find the following lines:

<path name="backuppath"> <param name="tapePath">/dev/rmt/0</param> </path>

Replace /dev/rmt/0 with the path of the tape drive in the system.

NOTE

the tape from the tape drive. After you run the command mt stat, you can determine the path of the tapedrive according to the system output.

If the system displays the following message, then you can determine that the path of the tape drive is /dev/rmt/0./dev/rmt/0n: no tape loaded or drive offline

Step 6 Save and close the file.

Step 7 Restart the M2000 services.

# . /opt/OMC/svc_profile.sh

# start_svc

Step 8 Check whether all the M2000 services are started.

# svc_adm -cmd status

Check the value after Not Running : in the last line. If the value is 0, you can infer that allthe M2000 services are started.

----End

10.2.3 Backing Up M2000 Dynamic Data PeriodicallyThis task is performed to periodically back up the M2000 dynamic data.

Prerequisitel You have logged in to the M2000 client.

l You have relevant rights.

l The tape drive connects properly and holds a tape if you plan to back up data on a tape.

l The hard disk has sufficient free space if you plan to back up data on a hard disk.

Context

Generally, M2000 dynamic data is backed up periodically. The periodic backup of the dynamicdata is performed in full backup mode.

The dynamic data backup has no restriction on backup time. The backup can be performed duringthe system operation.




Procedure

Step 1 Choose Tools > System Backup > Periodic Backup Setting .The Periodic Backup dialog box is displayed.

Step 2 Click the Server tab. Select Enable Server Periodic Backup and set Execution Time and FullBackup Day, as shown in Figure 10-1.

Figure 10-1 Periodic Backup dialog box

Step 3 Click OK.

The system automatically performs periodic backup.

----End

10.2.4 Backing Up M2000 Dynamic Data ManuallyThis task is performed to manually back up the M2000 dynamic data.

Prerequisitel You have logged in to the M2000 client.

l You have relevant rights.

l The tape drive connects properly and holds a tape if you plan to back up data on a tape.

l The hard disk has sufficient free space if you plan to back up data on a hard disk.

ContextManual backup is required in special or urgent situations, such as the loss of backup tapes or thefailure of the M2000 system.

Manual backup can be full backup or incremental backup.



Dynamic data backup has no restriction on backup time. The backup can be performed duringthe system operation.

Procedure

Step 1 Choose Tools > System Backup > Backup Management .

The Backup Management dialog box is displayed.

Figure 10-2 Backup Management dialog box

Step 2 On the Server Backup List tab page, as shown in Figure 10-2, click Full Backup to performfull backup. Click Incremental Backup to perform incremental backup.

NOTE

Full backup must be performed before incremental backup is performed.

----End

10.2.5 Restoring M2000 Dynamic DataThis task is performed to restore M2000 dynamic data.

Prerequisitel You have logged in as the root user.




l You have obtained the backup files for restoring the M2000 dynamic data.

l The M2000 is operational.

ContextBased on the latest full backup files and incremental backup files of the M2000, restore theM2000 system. After you restore the M2000 system, all the contents in the backup file packageare restored. If some data, such as the NE performance data or alarm data, is not packed in thebackup file package, the M2000 starts the automatic synchronization function to obtain andhandle the data.

During the restoration of the dynamic data, the M2000 services are stopped and thus the NEperformance or alarm data cannot be handled.

Procedure

Step 1 Switch to /opt/OMC, which is the default installation path of the M2000 server software.

# cd /opt/OMC

Step 2 Stop the running M2000 services.


# stop_svc

Step 3 Skip this step and go to Step 4 if the backup files are stored on a hard disk. Copy the backupfiles to a hard disk if the backup files are stored on a tape.

# cd /

# tar xf /dev/rmt/0

CAUTIONl The decompressed file must be stored in the same path as the compressed one. It is not related

to the path where you perform the decompression operation.l M2000 dynamic data cannot be restored if neither a tape nor a hard disk stores the backup

log file backup.log.

Step 4 Switch to user dbuser and restart the Sybase. Then switch to user root.

To know how to stop the Sybase, refer to 16.2.3 How to Stop the Sybase.

To know how to start the Sybase, refer to 16.2.2 How to Start the Sybase.

Step 5 Restore the dynamic data.

# svc_backuprestore -cmd restore

The following message is displayed:

****************************************************** * Restore is a dangerous operation ! * * If you have to do it, please stop system first ! * ******************************************************



Continue ?[n/y]

Enter y and start the restoration.

After the restoration is complete, the system displays the message restoresuccessfully!

Step 6 Restart the M2000 services.

# start_svc

----End

10.3 Backing Up and Restoring M2000 Static DataThis task is performed to back up and restore the M2000 static data.


10.3.2 Backing Up M2000 Static DataThis task is performed to back up M2000 static data.

10.3.3 Restoring M2000 Static DataThis task is performed to restore the M2000 static data.


The static data to be backed up involves the data in /opt/OMC, /opt/sybase, and /export/home ofthe hard disk on the M2000 server. The backup data is used for restoring the database or theM2000 program. Table 10-8 lists the content, partition, mode, and storage device of the backupof static data.

Table 10-8 Static data backup

Item Description

Content l M2000 server applications

l Database software

l Database files

l M2000 configuration files

Partition l /opt/OMC

l /opt/sybase

l /export/home




Item Description

Backup mode One-time backup mode. Generally, the dataneed not be backed up again after it is backedup for the first time. You, however, mustback up the data again if you upgrade adatabase program, expand the capability ofthe database, or upgrade M2000 applicationprograms.


Amount of required tapes In general, two 72 GB tapes are required forthe backup of static data. In practice,determine the amount of required tapesaccording to the used space of the files in aspecific partition or a path.NOTE

l You can use the command du -sh /opt/OMCto view the used space of the files in the /opt/OMC path.

l You can use the command du -sh /opt/sybase to view the used space of the files inthe /opt/sybase path.

l You can use the command df -h /export/home to view the used space of the files inthe /export/home partition.

10.3.2 Backing Up M2000 Static DataThis task is performed to back up M2000 static data.

Prerequisitel The ufsdump solution is applicable to all the M2000 systems that are installed with the

Solaris 10 operating system.

l Prepare the required tapes for system data backup.

l The magnetic head of the tape drive is clean. If the tape drive is externally connected, ensurethat the connection between the tape drive and the M2000 server is proper.

l You have logged in as the root user.

Context

The static data backup is used for recovering the M2000 system in case the system incurs asevere damage, for example, the damage of the M2000 or Sybase application software or theloss of the database files.

NOTE

During the backup, the system prompts you to replace the fully loaded tape with a new tape. the fully loadedtape, label the tape, and then insert another tape to continue with the installation.



Procedure

Step 1 Restore the dynamic data.For details, refer to 10.2.4 Backing Up M2000 Dynamic Data Manually.

Step 2 Switch to the M2000 installation path.The default M2000 installation path is /opt/OMC/opt/OMC.

Step 3 If the services and the Sybase server are running, stop the M2000 services and the Sybase.


# stop_svc

# su - dbuser

NOTE

Run the svc_ps command to check whether the services stop. If there are no output, you can infer thatservices are stopped. If the output still exists, run the kill_svc command to stop services. Ensure that nooutput exists when you run the svc_ps command.

-bash-3.00$ isql -SSYB -Usa -PPassword of sa

1> shutdown SYB_BACKUP

2> go

1> shutdown

2> go

-bash-3.00$ exit

NOTE

Run the ps -ef | grep sybase | grep -v grep command to check whether the Sybase stops. If no output exists,you can infer that the Sybase is already stopped. If the output still exists, run the kill -9 PID of Sybasecommand to force the Sybase to stop.

Step 4 Insert a tape into the tape drive, and then run the following command to rewind the tape:

# mt -f /dev/rmt/0 rewind

Store the backup files in two tapes if one tape cannot hold the backup files.

Step 5 Delete extra files if the backup is not performed immediately after the initial installation of theM2000. Skip this step and go to Step 6 if the backup is performed immediately after the initialinstallation of the M2000.

NOTE

If the customer needs to reserve these files, move these files to another folder. Move them back when thebackup is complete.

l To delete the alarm dump files, run the following commands:# cd /export/home/omc/var/fm/dump# rm -R /export/home/omc/var/fm/dump/event/*# rm -R /export/home/omc/var/fm/dump/history/*

l To delete the core files, run the following commands:# cd /export/home/omc/var/logs




# rm core*l To delete the history trace files, run the following commands:

# cd /export/home/omc/var/logs/tracebak# rm iMAP*

Step 6 Back up the M2000 application.

# cd /opt

# ufsdump 0ucf /dev/rmt/0n OMC

Step 7 Back up the database application.

# ufsdump 0ucf /dev/rmt/0n sybase

Step 8 Back up the M2000 configuration files and the device files of the Sybase.

# cd /

# ufsdump 0ucf /dev/rmt/0n /export/home

NOTE

The device files of the Sybase are master.dat, sysprocs.dat, and tmp_dev. The three files are stored in thedirectory /export/home/sybdev.

Step 9 Check whether the static backup data on the tape is valid. If the static backup data is backed upon multiple tapes, check the tapes one by one.1. Rewind the tape.

# mt rewind2. The contents of the first record on the tape are displayed.

# ufsrestore tf /dev/rmt/0n3. The contents of the second record on the tape are displayed.

# ufsrestore tf /dev/rmt/0n4. The contents of the third record on the tape are displayed.

# ufsrestore tf /dev/rmt/0nIf error messages such as Media read error: I/O error are not displayed, you caninfer that the backup data is valid.

----End

10.3.3 Restoring M2000 Static DataThis task is performed to restore the M2000 static data.

Prerequisitel You have logged in as the root user.

l The M2000 and database files (such as the files in /opt/OMC and /export/home, and theSybase files in /opt/sybase) have been deleted from the system.

l The tape storing the backup static data has been inserted into the tape drive.



ContextIf the M2000 applications are lost or the disk storing the database raw device breaks down owingto severe system damage, restore the static data.

NOTE

If the static data is backed up on more than one tape, insert the first tape to start the restoration, and theninsert the subsequent tapes as prompted.If a message such as Warning is displayed during the restoration of the static data, you can ignore themessage.Warning: ./lost+found:File exists

ProcedureStep 1 If the M2000 services and the Sybase server are running, stop the M2000 services and the Sybase.


# stop_svc

# su - dbuser

NOTE

Run the svc_ps command to check whether the services stop. If there are no output, you can infer thatservices are stopped. If the output still exist, run the kill_svc command to stop services. Ensure that nooutput exists when you run the svc_ps command.



2> go

1> shutdown

2> go

-bash-3.00$ exit

NOTE

Run the ps -ef | grep sybase | grep -v grep command to check whether the Sybase stops. If no output exists,you can infer that the Sybase is already stopped. If the output still exists, run the kill -9 PID of Sybasecommand to force the Sybase to stop.

Step 2 Check whether the raw volume or the disk raw partition used by the data device functions well,run the following command:

NOTE

Before you rebuild the subscriber database, ensure that the raw volume or disk raw partition used by thedata device functions well.

l In the case of Netra 240 single-server system, the raw volume used by the data device is created afteryou use the PlatformTools to encapsulate and mirror the operating system.

l For other cases, you need to use the PlatformTools to separately create volumes.

1. To check whether the raw volume or the disk raw partition used by the data device isavailable and functions well, run the following command:

# vxprint




If the omcdb_data_lv001and omcdb_log_lv003 are displayed, you can infer thatthe raw volume exists. Each volume (v) in the system output contains the values ofSTATE and KSTATE. If the value of STATE is ACTIVE and the value of KSTATE isENABLED, infer that the raw volume or the disk raw partition used by the data devicefunctions well.

2. Check whether the exporthome volume on the disk array is attached to the /export/homepath.

# df -k

Filesystem kbytes used avail capacity Mounted on/dev/md/dsk/d100 17146994 10350431 6625094 61% //devices 0 0 0 0% /devicesctfs 0 0 0 0% /system/contractproc 0 0 0 0% /procmnttab 0 0 0 0% /etc/mnttabswap 9496440 1320 9495120 1% /etc/svc/volatileobjfs 0 0 0 0% /system/objectfd 0 0 0 0% /dev/fdswap 9495184 64 9495120 1% /tmpswap 9495224 104 9495120 1% /var/runswap 9495120 0 9495120 0% /dev/vx/dmpswap 9495120 0 9495120 0% /dev/vx/rdmp/dev/md/dsk/d215 1020815 3604 955963 1% /global/.devices/node@1/dev/md/dsk/d115 1020815 3602 955965 1% /global/.devices/node@2/dev/vx/dsk/ossdg/exporthome 144506594 2647322 140414207 2% /export/home

In the above system output, the /dev/vx/dsk/ossdg/exporthome disk device must be attachedto the /export/home path.

Step 3 Insert a tape into the tape drive.

Step 4 Check whether the static backup data on the tape is valid. If the static backup data is backed upon multiple tapes, check the tapes one by one.1. Rewind the tape.


# ufsrestore tf /dev/rmt/0n3. The contents of the second record on the tape are displayed.

# ufsrestore tf /dev/rmt/0n4. The contents of the third record on the tape are displayed.

# ufsrestore tf /dev/rmt/0nIf error messages such as Media read error: I/O error are not displayed, you caninfer that the backup data is valid.

Step 5 Rewind the tape.

If the static backup data is stored on multiple tapes, the tapes from the tape drive. Insert the firsttape on which the static backup data is stored, and then rewind the tape.


Step 6 Restore the M2000 application.

# cd /



# rm -R /opt/OMC

# ufsrestore rf /dev/rmt/0n

Step 7 Restore the Sybase application.

# rm -R /opt/sybase


Step 8 Restore the M2000 configuration files and the device files of the Sybase.

# cd /export/home


# rm restoresymtable

Step 9 Start the Sybase as the dbuser user

CAUTIONl You must start the Sybase as the dbuser user. In this case, the owner of a created database

device is the dbuser user, and you are able to start the Sybase as the dbuser user in the future.l During the startup of the Sybase, if the system displays the message indicating that the device

associated with the subscriber database cannot be found, you can ignore the message.

# su - dbuser

-bash-3.00$ cd /opt/sybase/ASE-15_0/install/

-bash-3.00$ ./startserver -f ./RUN_SYB

-bash-3.00$ nohup ./startserver -f ./RUN_SYB > /dev/null

Step 10 Check the database status.


1> sp_helpdb

2> go

If all the user database exists and the value of the status is not offline or notrecovered in each database, you do not need to re-establish the database. That is, you do notneed to perform Step 11 through Step 14. Perform Step 15 to restore the real-time data.

Switch back to log in as the root user.

1> exit

-bash-3.00$ exit

Step 11 Run the PlatformTool to rebuild an application database.1. Start the PlatformTools.

# cd /opt/PlatformTools




# ./setup

NOTE

After you start the PlatformTool, you can enter R on the system interface to return to the upper-levelmenu, or enter Q to exit the tool.

2. At the following prompt, enter 2.=============================================================Please choose the operation type1--Installation2--MaintenanceQ--Exit=============================================================Please make a choice :

3. At the following prompt, enter 3 to choose OMC database manipulations.=============================================================1--Backup OS2--Restore OS3--OMC database manipulationsR--ReturnQ--Exit=============================================================Please make a choice:

4. At the following prompt, enter 1 to choose Rebuild OMC database.=============================================================1--Rebuild OMC database2--Remove OMC databaseR--ReturnQ--Exit=============================================================Please make a choice:

5. At the following prompt, enter y to start rebuilding the subscriber database.=============================================================NOTE:

If you only want to remove user databases, don't choose this operation. Before you continue, you must have sybase system databases restored. Stop all M2000 services , and make sure sybase dataserver is online.

Do you want to continue? [y/n]

6. At the following prompt, enter the name of the database sever. The default name of theserver is SYB.Please input Database server [default:SYB]:

7. At the following prompt, enter the name of the database administrator, whose default nameis sa.Please input Database user name [default:sa]:

8. At the following prompt, enter the password of the database administrator. The defaultpassword of the database administrator is emsems.Please input Database password [default:emsems]:

If the system displays no error message and the system output contains the messageRebuilding user databases finished, you can infer that the restoration issuccessful.

9. At the following prompt, enter Q to exit the Platformtool.Please restart Sybase ASE to continue ...=============================================================1--Rebuild OMC database2--Remove OMC databaseR--Return



Q--Exit=============================================================Please make a choice:

Step 12 Check the database status.


1> sp_helpdb

2> go

If all the user database exists and the value of the status is not offline or notrecovered in each database, you can infer that the application database is successfullyestablished and you can proceed with the following steps. If the application database is notestablished, check whether the interface on which the script is running can be correctly displayed.If the interface on which the script is running does not function well, contact Huawei engineers.

Switch back to log in as the root user.

1> exit

-bash-3.00$ exit

Step 13 Stop the Sybase.

# su - dbuser

-bash-3.00$ isql -SSYB -Usa -Ppassword of sa


2> go

1> shutdown

2> go

1> exit

Step 14 Start the Sybase.

-bash-3.00$ cd /opt/sybase/ASE-15_0/install/

-bash-3.00$ ./startserver -f ./RUN_SYB

-bash-3.00$ nohup ./startserver -f ./RUN_SYB > /dev/null

Step 15 Restoring the dynamic data.

For details, see 10.2.5 Restoring M2000 Dynamic Data.

Step 16 Ensure that all the M2000 services are started.

# svc_adm -cmd status

Check the value after Not Running : in the last line. If the value is 0, you can infer that allthe M2000 services are started.

In this case, the data restoration is complete.

----End




10.4 Backing Up and Restoring M2000 System DataThis task is performed to back up and restore the M2000 system data. The M2000 provides twobackup and restore solutions for backing up and restoring the system: ufsdump and Veritas. Theufsdump solution backs up the data on the system disk to the tape drive using the Solaris systemcommands and scripts. The ufsdump solution is applicable to all server types. The Veritassolution backs up data through the backup and restore components (NetBackup and BMR)provided by the Symantec company. The Veritas solution is applicable to all server types. Fordetails, refer to the manuals related to the Veritas system backup solution.

Context

CAUTIONl The ufsdump solution is not applicable to the case in which the operating system is installed

on multiple disks.l The system backup must be performed in single-user mode.


10.4.2 Backing Up M2000 System DataThis task is performed to back up the M2000 system data in ufsdump mode.

10.4.3 Restoring M2000 System Data (Solaris Volume Manager)This task is performed to restore the M2000 system data in ufsrestore mode in Solaris VolumeManager.

10.4.4 Restoring M2000 System Data (Veritas Volume Manager)This task is performed to restore the M2000 system data in ufsrestore mode in Veritas VolumeManager.


The M2000 system data is backed up for restoring the existing operating system, databasesoftware, and M2O00 server applications. Table 10-9 lists the content, partition, mode, andstorage device of the backup of system data.

Table 10-9 System data backup

Item Description

Content l Solaris

l Database software

l M2000 server application software

Partition /



Item Description

Backup mode One-time backup. The system data is backed up once forall unless a new operating system patch is installed.


Amount of required tapes In general, a 72 GB tape is required for the backup ofsystem data. In practice, determine the amount of requiredtapes according to the used space of the files in a specificpartition.NOTE

You can use the command df -h / to view the used space of thefiles in the / partition.

10.4.2 Backing Up M2000 System DataThis task is performed to back up the M2000 system data in ufsdump mode.

PrerequisiteThe ufsdump solution is used to back up and restore the data on the local system disk, ratherthan the data on disk arrays. The ufsdump solution used for the system backup has the followingrequirements:l The solution is applicable to all the M2000 systems that are installed with Solaris 10.

l Prepare the required tapes for system data backup.

l The magnetic head of the tape drive is clean. If the tape drive is externally connected, ensurethat the connection between the tape drive and the M2000 server is proper.

l The backup of the M2000 static data is complete. The M2000 services and the Sybase arestopped.

Procedure

Step 1 Check the servers before the backup.1. To check the physical and logical IP addresses of the servers, run the following command:

# ifconfig -a

Ensure that the current local IP addresses is set in consistency with the previous planning.2. Check the system hardware.

a. Check the LED indicators of all the system hardware parts to ensure that no LEDindicator is flashing for alarms.

b. Ensure that both the local disk and the disk array are included.Run the following command to check the screen output of the information about thelocal disk and the disk array:# format

c. Ensure that no hardware part is damaged.Run the following command to check that the screen output does not include theinformation about hardware damage, such as error or failed:




# prtdiag -v

3. Check whether the root disks are properly encapsulated and mirrored.l For root disks encapsulated with Solaris Volume Manager, run the following

commands:# metastat

In the system out, check that each Mirror has two Submirrors and that all the valuesof State are Okay.

l For root disks encapsulated with Veritas Volume Manager, run the followingcommands:# vxprint -th -g Name of disk group(rootdg)

In the displayed result, ensure that the values in the KSTATE column of the pl and vare ENABLED, and the values in the STATE column of the pl and v are ACTIVE.

4. Check the operating system logs.

Check the log file /var/adm/messages of the operating system. Ensure that no error logsare reported recently.

5. Run the explorer program.

The program explorer collects all hardware information of the existing system and thensaves the collected information to a package. To save the system information before theoperating system is backed up, you need to run the program explorer to generate theinformation about the system state.

To run the explorer program, run the following command:

# /opt/SUNWexplo/bin/explorer

The results generated after runing the explorer program are saved in the /opt/SUNWexplo/output directory. When you run the backup script later, the previous result file is backedup to a tape.

The following file is an example of the result file generated after running explorer:

explorer.83af631b.Athene-2006.01.21.10.01.tar.gz

NOTE

The file name contains the date and time when this file is generated.

Step 2 Back up the operating system in ufsdump mode.

1. Log in to the SC or the RSC of the server.To know how to log in to the system controller (SC) or the remote system controller (RSC)of the server, refer to the software installation manual related to each server type.

2. Stop the operating system.

# sync; sync; sync; sync; sync; sync

# /usr/sbin/shutdown -i0 -g0 -y

3. At the ok state, restart the operating system and change to the single-user mode.

ok boot -s

At the following prompt, enter the password of the root user.



Root password for system maintenance (control-d to bypass):4. Check the status of the tape drive.

After a system restart, insert a tape into the tape drive whose node requires backup.

Run the following command:

# mt stat

The state of the tape is normal if the command results are similar to the followinginformation:

HP DAT-72 tape drive: sense key(0x6)= Unit Attention residual= 0 retries= 0 file no= 0 block no= 0

If a tape is not mounted or identified correctly, the following information is displayed:

/dev/rmt/0n: no tape loaded or drive offline

Before continuing the operation, you must enable the tape drive to identify a tape.

Step 3 Use the PlatformTools to back up the system data.1. Start the PlatformTools.


# ./setup

NOTE


2. At the following prompt, enter 2.=============================================================Please choose the operation type1--Installation2--MaintenanceQ--Exit=============================================================Please make a choice:

3. At the following prompt, enter 1.=============================================================1--Backup OS2--Restore OS3--OMC database manipulationsR--ReturnQ--Exit=============================================================Please make a choice:

4. At the following prompt, enter y.=============================================================NOTE:

For Solaris 10 system only. Boot disk should be encapsulted and mirrored by VxVM or SVM.

Do you want to continue? [y/n]

After you enter y, the script completes the system backup automatically, collect statisticson the total time taken for the system backup, and displays the time. The system backuptakes about one hour.




The contents to be backed up on the tape are as follows:l PlatformTools

l Data of root disks

l Restore script

l Result of executing the explorer program

If the system displays no error message and the system output contains the messageSuccessfully completed operating system backup!, you can infer thatthe backup of the operating system is successful.

Step 4 Exit the single-user mode, and then return to the multiuser mode.

# exit

After the steps mentioned earlier are performed, the system backup is complete. the tape, andthen attach a label to it. The label contains the backup contents and the date.

Step 5 Check whether the system backup data on the tape is valid.1. Rewind the tape.

# mt rewind

2. The contents of the first record on the tape are displayed.

# tar tf /dev/rmt/0n

3. The contents of the second record on the tape are displayed.

# ufsrestore tf /dev/rmt/0n

If error messages such as Media read error: I/O error are not displayed, you caninfer that the system backup data is valid.

----End

10.4.3 Restoring M2000 System Data (Solaris Volume Manager)This task is performed to restore the M2000 system data in ufsrestore mode in Solaris VolumeManager.

ContextWhen the PlatformTools is used to restore the operating system, as described in Step 8, the toolautomatically determines the encapsulation type of the operating system.

Procedure

Step 1 Log in to the system controller (SC) or remote system controller (RSC) of the server.To know how to log in to the SC or the RSC, refer to the related M2000 software installationguide.

Step 2 Power off the disk array

If the server has disk arrays, such as 3320 (3310) and 6140 (6130), power off the disk arrays. Ifthe server has no disk arrays, go to the next step.



CAUTIONl If you boot the system from the CD-ROM, power off the disk arrays first.

l Ensure that the logical serial numbers of the local disks that are booted from the CD-ROMare identical to those in the existing system.

Step 3 Start the operating system using a disk.Insert the Solaris installation disk into the CD-ROM drive of the server to be restored, and followthe guides displayed by the CD-ROM to start the operating system for the restoration of thesystem data. If the operating system is started, enter /usr/sbin/shutdown -g0 -y -i0 to switch tothe state of the ok prompt. If the operating system is in the state of the ok prompt, run thefollowing command:

okboot cdrom -s

Step 4 Check the state of the tape drive.

After the system starts, insert a tape into the tape drive of the node that requires the backupoperation.


# mt stat

If the command results are similar to the following information, you can infer that the insertedtape is functioning well:


If you do not insert the tape into the tape drive correctly or the tape drive cannot identify thetape, the following message is displayed:


Handle the problem before you perform the backup operation


# mt rewind

2. The contents of the first record on the tape are displayed.

# tar tf /dev/rmt/0n

3. The contents of the second record on the tape are displayed.

# ufsrestore tf /dev/rmt/0n

If error messages such as Media read error: I/O error are not displayed, you caninfer that the system backup data is valid.


If the system backup data is stored on multiple tapes, the tapes from the tape drive. Insert thefirst tape on which the system backup data is stored, and then rewind the tape.





Step 7 Restore the PlatformTools from the tape drive to the directory /tmp.

# tar xf /dev/rmt/0n

Step 8 Start the PlatformTools to automatically restore the system.1. Start the PlatformTools.

# cd /tmp/PlatformTools

# ./setup

NOTE



3. At the following prompt, enter 2.=============================================================1--Backup OS2--Restore OS3--OMC database manipulationsR--ReturnQ--Exit=============================================================


For Solaris 10 system only. The server must be booted into single user mode from cdrom to do restoring.

Do you want to continue? [y/n] y

The tool determines the original encapsulation type of the system disk and automaticallysearches for the original disk that is installed with the operating system to restore the systemdata.l At the following message, you can infer that the encapsulation type is Solaris Volume

Manager. After you perform Step 8.5, go to Step 9.Checking original boot disk encapsulation type...The original boot disk is Solaris volume manager encapsulated.

l At the following message, you can infer that the encapsulation type is Veritas VolumeManager, after you perform Step 8.5, go to Step 9 in 10.4.4 Restoring M2000 SystemData (Veritas Volume Manager).Checking original boot disk encapsulation type...The original boot disk is Veritas volume manager encapsulated.

At the prompt Is application data on disk array(y/n)?,

l If the server type is Netra 240, enter n. The Netra 240 sever has no disk arrays and itsapplication data is stored on the disk where the operating system is installed.



l If the server type is V890 or E4900, enter y. The V890 or E4900 server has disk arraysand its application data is stored on the disk arrays.

After you press Enter, the script automatically partitions root disks, creates a file system,restores the partition data of the root disks, rebuilds a device tree, creates a guide block,and modifies the system and vfstab files to unpack the encapsulation of Volume Manager.

If the system displays no error message and the system output contains the messageSuccessfully completed operating system restore!, you can infer thatthe restoration of the operating system is successful.

5. After the system data is successfully restored, run the following command to stop theoperating system:

# sync; sync; sync; sync; sync; sync; reboot -- -r

Step 9 If the system disk is encapsulated through Solaris Volume Manager, perform the following stepsto re-encapsulate the system disk after the operating system is restarted:After the operating system is restored, use the PlatformTools tool to re-encapsulate the systemdisk.1. Start the PlatformTools tool.


# ./setup

NOTE

After the PlatformTool tool is running, you can enter R to return to the upper-level menu , or enterQ to quit.

2. At the following prompt, enter 1 to choose Installation.=============================================================Please choose the operation type1--Installation2--MaintenanceQ--Exit=============================================================Please make a choice:

3. At the following prompt, enter 1 to choose the M2000 system as the single-server system.=============================================================Please choose the system type 1--Single system 2--Dual system Q--Exit

=============================================================

4. At the following prompt, choose the server type based on your requirement.

NOTE

The parameter 3300Array represents the disks arrays in the 3300 series, such as 3310 and 3320.=============================================================Please choose the machine type1--E4900+61302--E4900+61403--V890+3300Array4--Netra240R--ReturnQ--Exit=============================================================Please make a choice :




System Configuration Enter

Sun E4900 server + 6130 disk array 1


Sun V890 server + disk arrays in the 3300 series 3

Sun Netra240 server 4

5. At the following prompt, enter 2 to choose Encapsulate local disks.=============================================================1--Customize the Solaris 2--Encapsulate local disks 3--Mirror local disks 4--Create volumes 5--Set the parameters of Sybase 6--Create database of OMC R--Return Q--Exit =============================================================Please make a choice:

6. At the following prompt, enter y to start the encapsulation.Do you want to encapsulate the local disks? [y/n]

At the following prompt, which indicates that the encapsulation is complete, press Enter.

Encapsulate the local disks completed successfully Hit RETURN to go back

Step 10 If the server has disk arrays, turn on the power of the disk arrays, refresh the hardwareinformation, compile the file system list, namely the file /etc/vfstab, and add the attachment pointinformation on the array directory after the local disks are encapsulated. If the sever has no diskarrays, go to the next step.1. Turn on the power of the disk arrays. Wait for about five minutes.2. Search for the hardware.

# devfsadm3. Check whether the system output contains the information on both local disks and disk

arrays.

# format

NOTE

If you confirm that the system output is correct, press Ctrl+D to return to the state of the commandline.

4. Refresh the disk information that is stored in the Veritas volume management software.

# vxdctl enable5. Check whether the system output contains the information on both local disks and disk

arrays.

# vxdisk list6. Set the terminal type.

# TERM=vt100; export TERM7. Run the following command:

# vi /etc/vfstab



Append the following line to the file:

/dev/vx/dsk/ossdg/exporthome /dev/vx/rdsk/ossdg/exporthome /export/home ufs 2 yes logging

NOTE

When adding the previous line, you are not allowed to press Enter to change to the next line. Inaddition, you must press Tab to keep the space between fields. If you press Enter to change to thenext line, or press another key except Tab to keep the space between fields, errors may occur.

Step 11 Restart the operating system.


NOTE

l You must use the command sync; sync; sync; sync; sync; sync; reboot -- -r.

l Two consecutive symbols "-" exist between reboot and -r.

l If the operating system cannot be started, switch back to the ok state, and then run the commandboot -ar to restart the operating system again. At the system prompts, press Enter repeatedly until theoperating system is successfully started.

Step 12 Mirror the root disk.NOTE

l After the operating system is restored, the mirror settings of the previous local disks become invalidand you have to remirror the local disks.

l The mirrors of the disk arrays are still valid.

After the system is restarted and switched to the cluster mode, perform the following steps tomirror the root disks.1. Start the PlatformTools tool.


# ./setup

NOTE

After the PlatformTool tool is running, you can enter R to return to the upper-level menu or enterQ to quit.

2. At the following prompt, enter 1 to choose Installation.=============================================================Please choose the operation type1--Installation2--MaintenanceQ--Exit=============================================================Please make a choice:

3. At the following prompt, enter 1 to choose the single-server system.=============================================================Please choose the system type 1--Single system 2--Dual system Q--Exit

=============================================================4. At the following prompt, choose the type of the M2000 server based on your requirement.

NOTE

The parameter 3300Array represents the disks arrays in the 3300 series, such as 3310 and 3320.




=============================================================Please choose the machine type 1--E4900+6130 2--E4900+6140 3--V890+3300Array 4--Netra240 R--Return Q--Exit =============================================================Please make a choice:

System Configuration Enter



Sun V890 server + disk arrays in the 3300 series 3

Sun Netra240 server 4

5. At the following prompt, enter 3 to choose Mirror local disks.=============================================================1--Customize the Solaris 2--Encapsulate local disks 3--Mirror local disks 4--Create volumes 5--Set the parameters of Sybase 6--Create database of OMC R--Return Q--Exit =============================================================Please make a choice:

6. At the following prompt, enter y to start the mirroring.Do you want to mirror the local disks? [y/n]

The time used for mirroring varies with the server configuration. It takes about 1-3 hours.During the mirroring, the system displays the following message to tell you thesynchronization progress:

Please wait ...Synchronization progress: 0 % completeSynchronization progress: 0 % complete...Synchronization progress: 0 % complete...

When the system displays the following message, which indicates that the mirroring iscomplete, press Enter.

Mirror the local disks completed successfully.Hit RETURN to go back

7. At the following prompt, enter Q to exit the Platformtool.=============================================================1--Customize the Solaris 2--Encapsulate local disks 3--Mirror local disks 4--Create volumes 5--Set the parameters of Sybase 6--Create database of OMC R--Return Q--Exit =============================================================Please make a choice:



Step 13 Restore the M2000 static data.To know how to restore the M2000 static data, refer to 10.3.3 Restoring M2000 Static Data.

----End

10.4.4 Restoring M2000 System Data (Veritas Volume Manager)This task is performed to restore the M2000 system data in ufsrestore mode in Veritas VolumeManager.

ContextWhen the PlatformTools is used to automatically restore the operating system, as described inStep 8, the tool automatically determines the encapsulation type of the operating system.

Procedure

Step 1 Log in to the system controller (SC) or remote system controller (RSC) of the server.To know how to log in to the SC or the RSC, refer to the related M2000 software installationguide.

Step 2 Power off the disk array

If the server has disk arrays, such as 3320 (3310) and 6140 (6130), power off the disk arrays. Ifthe server has no disk arrays, go to the next step.

CAUTIONl If you boot the system from the CD-ROM, power off the disk arrays first.

l Ensure that the logical serial numbers of the local disks that are booted from the CD-ROMare identical to those in the existing system.

Step 3 Start the operating system using a disk.Insert the Solaris installation disk into the CD-ROM drive of the server to be restored, and followthe guides displayed by the CD-ROM to start the operating system for the restoration of thesystem data. If the operating system is started, enter /usr/sbin/shutdown -g0 -y -i0 to switch tothe state of the ok prompt. If the operating system is in the state of the ok prompt, run thefollowing command:

okboot cdrom -s

Step 4 Check the state of the tape drive.

After the system starts, insert a tape into the tape drive of the node that requires the backupoperation.


# mt stat

If the command results are similar to the following information, you can infer that the insertedtape is functioning well:





If you do not insert the tape into the tape drive correctly or the tape drive cannot identify thetape, the following message is displayed:


Handle the problem before you perform the backup operation.



# tar tf /dev/rmt/0n3. The contents of the second record on the tape are displayed.

# ufsrestore tf /dev/rmt/0nIf error messages such as Media read error: I/O error are not displayed, you caninfer that the system backup data is valid.


If the system backup data is stored on multiple tapes, the tapes from the tape drive. Insert thefirst tape on which the system backup data is stored, and then rewind the tape.


Step 7 Restore the PlatformTools from the tape drive to the directory /tmp.

# tar xf /dev/rmt/0n

Step 8 Start the PlatformTools to automatically restore the system.1. Start the PlatformTools.

# cd /tmp/PlatformTools

# ./setup

NOTE



3. At the following prompt, enter 2.=============================================================1--Backup OS2--Restore OS3--OMC database manipulationsR--Return



Q--Exit=============================================================


For Solaris 10 system only. The server must be booted into single user mode from cdrom to do restoring.

Do you want to continue? [y/n] y

The tool determines the original encapsulation type of the system disk and automaticallysearches for the original disk that is installed with the operating system to restore the systemdata.l At the following message, you can infer that the encapsulation type is Veritas Volume

Manager. After you perform Step 8.5, go to Step 9.Checking original boot disk encapsulation type...The original boot disk is Veritas volume manager encapsulated.

l At the following message, you can infer that the encapsulation type is Solaris VolumeManager. After you perform Step 8.5, go to 10.4.3 Restoring M2000 System Data(Solaris Volume Manager) in Step 9.Checking original boot disk encapsulation type...The original boot disk is Solaris volume manager encapsulated.

At the prompt Is application data on disk array(y/n)?,

l If the server type is Netra 240, enter n. The Netra 240 sever has no disk arrays and itsapplication data is stored on the disk where the operating system is installed.

l If the server type is V890 or E4900, enter y. The V890 or E4900 server has disk arraysand its application data is stored on the disk arrays.

After you press Enter, the script automatically partitions root disks, creates a file system,restores the partition data of the root disks, rebuilds a device tree, creates a guide block,and modifies the system and vfstab files to unpack the encapsulation of Volume Manager.

If the system displays no error message and the system output contains the messageSuccessfully completed operating system restore!, you can infer thatthe restoration of the operating system is successful.

5. After the system data is successfully restored, run the following command to stop theoperating system:


NOTE



Step 9 Delete the remaining rootdg message.

# vxdg destroy rootdg

Step 10 If the system disk is encapsulated through Veritas Volume Manager, perform the following stepsto re-encapsulate the system disk after the operating system is started:1. Start encapsulating the M2000 system disk.

# vxdiskadm




2. At the following prompt, enter 2 to select the menu Encapsulate one or moredisks, and then press Enter.Select an operation to perform:

3. At the following prompt, enter list to list the disk information.Select disk devices to encapsulate: [<pattern-list>,all,list,q,?]The system shows the following disk information:DEVICE DISK GROUP STATUS c0t0d0 - - online invalid c0t1d0 - - online invalid

4. At the following prompt, enter the device number of the system disk. Enter the devicenumber of the system disk, for example, c1t0d0.Select disk devices to encapsulate: [<pattern-list>,all,list,q,?]The system shows the following information: Here is the disk selected. Output format: [Device_Name]

c1t0d05. At the following prompt, enter y.

Continue operation? [y,n,q,?] (default: y)6. At the following prompt, enter rootdg.

Which disk group [<group>,list,q,?]7. At the following prompt, enter y.

Create a new group named rootdg? [y,n,q,?] (default: y)8. At the following prompt, enter n.

Use a default disk name for the disk? [y,n,q,?] (default: y)9. At the following prompt, enter y.

Continue with operation? [y,n,q,?] (default: y)10. At the following prompt, enter y.

Continue with encapsulation? [y,n,q,?] (default: y)11. At the following prompt, enter rootdg_1.

Enter disk name for c0t0d0 [<name>,q,?] (default: rootdg01)12. At the following prompt, press Enter.

Enter desired private region length [<privlen>,q,?] (default: 2048)13. At the following prompt, enter y to start encapsulate the root disk.

Encapsulate other disks? [y,n,q,?] (default: n)14. At the following prompt, enter q to complete the encapsulation.

Select an operation to perform:The system displays Goodbye..

Step 11 Prohibit DMP.1. Start prohibiting DMP.

# vxdiskadm2. At the following prompt, enter 17 to select Prevent multipathing/Suppress

devices from VxVM's view.Select an operation to perform:

3. At the following prompt, press Enter.Do you want to continue ? [y,n,q,?] (default: y)

4. At the following prompt, enter 5 to select Prevent multipathing of all diskson a controller by VxVM.



Select an operation to perform:

5. At the following prompt, enter all.Enter a controller name [<ctlr-name>,all,list,list-exclude,q,?]

6. At the following prompt, enter y.Continue operation? [y,n,q,?]

7. At the following prompt, press Enter.Hit RETURN to continue.

8. At the following prompt, enter q, and then press Enter to return to the upper menu.Select an operation to perform:

9. At the following prompt, enter q, and then press Enter to exit.Select an operation to perform:

Step 12 Turn on the power of disk arrays and refresh device information.

If the sever has disk arrays, perform the following steps to refresh hardware information afterthe operating system is started. If the sever has no disk arrays, go to the next step.

1. Power on the disk arrays. Wait for about five minutes.2. Search for the hardware.

# devfsadm

3. Check whether the system output contains the information on both local disks and diskarrays.

# format

NOTE

If you confirm that the system output is correct, press Ctrl+D to return to the command-line state.

4. Refresh the disk information that is stored in the Veritas volume management software.

# vxdctl enable

5. Check whether the system output contains the information on both local disks and diskarrays.

# vxdisk list

Step 13 Restart the operating system.


NOTE



l If the operating system cannot be started, switch back to the ok state, and then run the commandboot -ar to restart the operating system again. At the system prompts, press Enter repeatedly until theoperating system is successfully started.

Step 14 Check the encapsulation.

# df -k

In the system output, if the Filesystem columns of the /, /metadb, and /globaldevicespartitions start with /dev/vx/dsk/, you can infer that the system disk is successfully encapsulated.




Step 15 Mirror the M2000 local disks.

NOTE

The conventions for a root disk and a mirror disk are as follows: A root disk refer to the disk where theoperating system is installed before the system disk is mirrored; a mirror disk refers to the disk that is usedfor mirroring the system disk.

1. List disk information. The c1t0d0 is used for installing Solaris and c1t1d0 for mirroringthe system disk in this example.

# vxdisk list

DEVICE TYPE DISK GROUP STATUS c1t0d0s2 sliced rootdg_1 rootdg online c1t1d0s2 sliced - - online invalid c2t0d0s2 auto:cdsdisk - - online invalid

NOTE

l The DEVICE information displayed here varies according to the root disk configured on theserver.

l If status of c1t1d0 is online, run the following command to uninstall this disk from theVxVM:

# vxdiskunsetup -C name of the mirror disk (for example, c1t1d0)

1. Initialize the mirror disk.

# vxdctl enable

# vxdisksetup -if name of the mirror disk (for example, c1t1d0)2. Add the mirror disk into the disk group rootdg.

# vxdg -g name of the disk group (rootdg) adddisk rootmirr_1=name of the mirror disk(for example, c1t1d0s2)

3. Check whether the mirror disk is added to the disk group roodg.

# vxdisk -g rootdg list

DEVICE TYPE DISK GROUP STATUS c1t0d0s2 sliced rootdg_1 rootdg online c1t1d0s2 sliced rootmirr_1 rootdg online

In the system output, if status of the c1t1d0 is online, you can infer that the mirrordisk c1t1d0 is successfully added to the rootdg.

4. Mirror the system disk.

# vxmirror -g name of the disk group (rootdg) rootdg_1 rootmirr_1

NOTE

l The mirroring process lasts for about one hour.

l When the mirroring is complete, run vxtask list to check the synchronization status. Ensure thatthe system is not restarted before the synchronization is complete.

5. Check the mirror of the root disk.

# vxprint -g name of the disk group (rootdg)

TY NAME ASSOC KSTATE LENGTH PLOFFS STATE TUTIL0 PUTIL0 dg rootdg rootdg - - - - - -

dm rootdg_1 c1t0d0s2 - 286657920 - - - - dm rootmirr c1t1d0s2 - 286657920 - - - -



v globaldevices fsgen ENABLED 2116608 - ACTIVE - - pl globaldevices-01 globaldevices ENABLED 2116608 - ACTIVE - - sd rootdg_1-04 globaldevices-01 ENABLED 2116608 0 - - - pl globaldevices-02 globaldevices ENABLED 2116608 - ACTIVE - - sd rootmirr-03 globaldevices-02 ENABLED 2116608 0 - - -

v metadb fsgen ENABLED 427392 - ACTIVE - - pl metadb-01 metadb ENABLED 427392 - ACTIVE - - sd rootdg_1-03 metadb-01 ENABLED 427392 0 - - - pl metadb-02 metadb ENABLED 427392 - ACTIVE - - sd rootmirr-04 metadb-02 ENABLED 427392 0 - - -

v rootvol root ENABLED 225296640 - ACTIVE - - pl rootvol-01 rootvol ENABLED 225296640 - ACTIVE - - sd rootdg_1-B0 rootvol-01 ENABLED 1 0 - - Block0 sd rootdg_1-02 rootvol-01 ENABLED 225296639 1 - - - pl rootvol-02 rootvol ENABLED 225296640 - ACTIVE - - sd rootmirr-01 rootvol-02 ENABLED 225296640 0 - - -

v swapvol swap ENABLED 4111104 - ACTIVE - - pl swapvol-01 swapvol ENABLED 4111104 - ACTIVE - - sd rootdg_1-01 swapvol-01 ENABLED 4111104 0 - - - pl swapvol-02 swapvol ENABLED 4111104 - ACTIVE - - sd rootmirr-02 swapvol-02 ENABLED 4111104 0 - - -

In the system output, the mirroring is successful if the STATE of each v and pl isACTIVE.

Step 16 Restore the M2000 static data.To know how to restore the M2000 static data, refer to 10.3.3 Restoring M2000 Static Data.

----End




01-10 Backing Up and Restoring the M2000

Documents