Sources

Actions:  

Edit: You may edit parameters of allowed fields in any of the existing Source record.  

Clone: You may clone any existing Source record to create a new one with few unique parameters.  

Checkhost: Shows the verification status of the Source. Adding a Source triggers a verification of the Amanda client/directory combination. Depending on the result of the last verification, the icon will change to a green check mark (all systems go), red dash sign (error), a crescent for partial (meaning the source verification is not success on any of the backupsets) a question mark (?) (meaning the entry has not been checked yet),.  

You can also check all of the entries at once by clicking the Check All Hosts button at the bottom of the page. The message box displays the status of the Check All Hosts process: when the check is completed, and whether any errors were found.  

Activate/De-activate: You may toggle between activating and deactivating the Sources from the associated backup sets.  

Delete: You may delete any/all of the existing Source records.  

Types of Sources:  

Supported object types include the following. All objects are licensed. Note that the ZMC indicates how many licenses have been purchased and how many remain available for those object types that require licensing. Application Agents are described in more specific detail here.   

Click the Add Source button at the top of the page to create a new entry, or select an entry from the table to edit. You can also clone an existing entry with a different Name and path/hostname to create a new source.  

 

 

 

 

File systems:  

  

Linux /UNIX/ Mac OSX File System  

Lets you select a Linux, UNIX, or Mac OSX hostname and directory path for backup. Also lets you specify files to exclude, and whether encryption and compression should be used. Amanda does not cross file system boundaries and thus each file system/partition should be entered as its own entry (separate Source object).  

  

NDMP  

Lets you perform a backup of NDMP appliances/filers using NDMPv4 protocol. Netapp filers, Sun Unified Storage and BlueArc Titan Storage are supported. For further details on NDMP backups, see "NDMP appliances".  

  

NFS/iSCSI/Lustre  

Lets you perform a backup of NFS/iSCSI/Lustre protocols on the specified mounted directory path. For further details on NFS/iSCSI/Lustre backups.

  

Network/ CIFS File System  

Lets you select a Common Internet File System by hostname, share name, and domain (you will need the username and password to access the share). For further details on CIFS backups, see "Backing Up and Restoring Common Internet Filesystem Shares."  

  

Solaris File System  

Lets you select a Solaris file system by hostname and directory path for backup. Also lets you specify files to exclude, and whether encryption and compression should be used.  For further details, see "Solaris client".  

  

Windows File System (NTFS/ReFS)  

Lets you select a Windows file system for backup, and enable compression if desired.  

  

Windows System State  

Allows you to back up the MS Windows System State, and to enable compression as desired.  

  

Windows Template  

Lets you select a template that you create on the Windows backup client that defines what is to be backed up. See Using the Zmanda Windows Client Configuration Utility for details on template creation. It also allows you enable compression as desired.  

  

Databases:  

  

Microsoft SQL Server  

Lets you select a Windows SQL server by hostname for backup, and whether compression should be used. For further details, see "Backing Up and Restoring Microsoft SQL Servers."  

  

Oracle on Window  

Lets you select a Windows Oracle server by hostname for backup, and whether compression should be used. For further details, see "Backing Up and Restoring Oracle Servers (Windows)."  

  

Oracle on Linux/Solaris  

Lets you select an Linux or Solaris Oracle server for backup, and to enable compression if desired. You must also specify an SID List Name for the Oracle database. For further details, see "Backing Up and Restoring Oracle Servers (Linux/Solaris)."  

  

PostgreSQL  

Lets you select a PostgreSQL database for backup by specifying a hostname and data directory, and to enable encryption and compression if desired. For further details, see "Backing Up and Restoring PostgreSQL Servers."  

  

Applications:  

  

Microsoft Exchange  

Lets you select a Windows Exchange server by hostname for backup, and whether compression should be used. For further details, see "Backing Up and Restoring Microsoft Exchange Servers."  

  

Microsoft Sharepoint  

Lets you select a host running Microsoft Sharepoint 2007 server or WSS 3.0 server to be backed up. For further details, see "Backing Up and Restoring Microsoft Sharepoint Servers."  

  

HyperVisor:  

  

Microsoft Hyper-V Server:  

Lets you select a Microsoft Hyper-V server to backup guest VM images. For further details on Microsoft Hyper-V backups.  

  

VMWare ESX  

Lets you select a VMWare ESX server to backup guest VM images. For further details on VMware ESX backups, see "VMware Vsphere and ESX".  

  

Creating a Source:  

Click Add Source on the top of the page to create a new entry, or select an entry from the table to edit. You can also duplicate an existing entry (see below) and then edit the entry before using the Add button to add it to the table of backup objects.  

 

 

 

 

Source Type  

Source Type can be any of the source types described in the section above.  

  

Hostname  

Specify or select a hostname (or IP address) to back up. Note that a backup set cannot include duplicate objects (a.k.a. DLEs); if you attempt to add host/directory combination that already exists in the backup set, an error is displayed. For purposes of host and pathname collision detection, the characters:, \, and / are considered the same by the ZMC.  

  

Directory/Path  

Specify a directory on the currently selected host to back up. Unless excluded (see below), all directories/ files below this directory are recursively included in the backup. There is a limit of 255 characters for a directory name. Note that Amanda will not cross file system boundaries when completing backups on Linux filesystems.  For example, if \ is specified as the directory to back up, \tmp will not be included in the backup if it resides on a separate file system.  Solaris filesystems backup will cross filesystem boundaries. So it is important to exclude network directories (such as NFS or CIFS mounted directories).  

  

Encryption & Compression Options   

Lists Encryption and Compression choices. Encryption and Compression of data are described here.  

This section describes the compression and encryption options common to most of the backup object types.  

  

Data Deduplication  

Enabling Data Deduplication checkbox prevents backing up of duplicate data in the specified path, which include all directories and sub-directories.   

Built-in AES encryption and LZMA compression of the stored data happen by default. Hence, enabling data deduplication will disable the options for user to choose encryption and compression.   

Note: zbackup needs to be installed on the AE Server to enable Data Deduplication.  

  

Data Compression  

ZMC compresses the data on the Amanda server or client. Amanda client compression resulting in more efficient use of bandwidth when the backup image is sent to the server. ZMC supports compression using gzip, which creates archives that can be extracted almost universal. Compression levels - fast, best, and custom can be specified. The fast compression provides smaller backup window. The best compression likely provides a smaller backup sizes. The custom compression provides user an option to specify different compression command. Most common custom compression program used is /bin/pigz - this compression command provides faster, parallel compression that uses multi-core CPUs better.  

Many Tape drives have built-in hardware compression. There are many advantages in allowing such hardware to handle the compression task.  

TIP: Devices that use a proprietary compression can fail or otherwise becomes unusable, presenting difficulties in restoring from backups that were written to it. For images and other pre-compressed files, do not consume the backup window by pointlessly re-compressing them.  

  

Data Encryption  

Backup data should be secured as carefully as you would protect the live version. Encrypting backup data adds a layer of protection against misuse.  

Backup encryption can be performed on the server or on the client.  It is important to store encryption key passphrases and certificates securely. Backups cannot be retrieved if the passphrase files or certificates are lost.   

Encryption passphrase or keys must be provided during the restore process for the backup images to be decrypted. Amanda Enterprise does not provide encryption key management. Customers should make sure the encryption keys are backed up.  

Note: Encryption is a CPU intensive task. Enable it with care.  

  

Linux/Solaris/Mac clients  

ZMC encrypts the data on the Amanda server and Linux/Solaris clients using the amcryptsimple program, which uses Gnu Privacy Guard (GPG) to perform symmetric data encryption.  

Encryption passphrases are stored in the amandabackup user directory on the server. It is important to keep the encryption passphrases (default passphrase - /var/lib/amanda/.am_passphrase) safely and securely. The data cannot be restored without the passphrase. It is important to backup the passphrases on a regular basis by adding /var/lib/amanda directory as an Amanda DLE without enabling encryption for that DLE.  Also, keep a backup of the passphrase in another secure location (for example: printed hardcopy).  

ZMC uses encryption passphrase from the Amanda server to do restoration of Linux/Solaris/Mac encrypted backup images. You should copy the appropriate encryption passphrase file to the Amanda server - /var/lib/amanda/.am_passphrase (Linux/Mac) or /opt/zmanda/amanda/amanda/.am_passphrase (Solaris) before doing restoration of client encrypted backups using ZMC.   

You can use the command line tool - amrecover running on the Amanda client to restore client encrypted backups. The command line tool uses the encryption passphrase from the Amanda client to decrypt client encrypted backup images.  

  

Windows clients:  

Backup encryption on Windows client is performed using AES encryption keys. Please see Windows Client manual for more details.  

  

Exclude Paths:  

Lists Files to be excluded from the backup. Space-separated, quoted shell/tar glob expressions (i.e. * and ? wildcards) are allowed to specify multiple files and paths. See Exclude Specifications for more details. For example: If you are backing up the root file system in Solaris, you can exclude   

"./platform" "./system" "./proc" "./tmp" "./dev" directories/file systems.  

  

Exclude Specifications:  

Excluding files can optimize the performance of the backup set, especially one that would otherwise back up an entire host from the root directory down.  

Exclude specifications depend on the object type.  The patterns supported are different for Linux/Solaris/Mac OS X and Windows. Please see the next two sections for the details.  

  

Linux/Solaris/Mac OS X filesystems - GNU tar (default)  

The Linux/Solaris/Mac OS X filesystems use the GNU-tar utility (unless extended attribute backup is enabled), which supports exclude patterns.  If a backup of extended attributes is enabled (schily tar is used), exclude pattern cannot be specified.  

The ZMC can accept one or more explicit pathnames or wildcard patterns per backup object/disk list entry, separated by a space. Some simple examples for GNU tar, Windows clients and Schily tar below.  

  

GNU tar  

• You can explicitly exclude any file or directory by pathname. For example, it is recommended that you avoid backing up staging areas for backup sets, so if you are backing up a root directory (/) that includes staging area /var/lib/amanda/staging/, the exclude specification would be 
  ./var/lib/amanda/staging. If the backup object/DLE is set to back up /var, the exclude specification would be ./lib/amanda/staging. The pathname in exclude specification should be relative to the DLE directory.  
• To exclude any file or directory that ends in ".log" e.g. ppp.log, and Xfree86.0.log, specify *.log  
•  To exclude any file or directory with the string "log" e.g. logfile, maillog, syslog, and ppp.log,XFree86.0.log, specify *log*  
•  To exclude any file or directory that starts with string "cron" and ends in ".gz" e.g. cron.1.gz, cron.2.gz, and log/cron.1.gz, specify *cron*.gz  
• The question mark can be used to specify a single character. e.g. to exclude log.1 and log.2, specify log.?  
• Multiple patterns are allowed, separated by spaces. For example, specifying following would exclude this.doc, that.doc, and everythingelse.doc, and would also exclude the Misc Files directory. The string must be quoted if it includes spaces (such as "Misc Files").  

"./*.doc" "./Misc Files"  

• Use the backslash to escape any double-quote characters included in the file or pathname itself. For example, specifying "*.sh" "foo bar" "some\"quote" excludes *.sh, foo bar, and some"quote.  
• The ZMC will save the exclude pattern in a standard format that includes explicit quotes (and escapes any characters that require it). For example, if you specify:  

exclude  "*.doc"  *.txt  

      is saved as  

exclude  "*.doc"  "*.txt"  

• Shadowed excludes are automatically deleted whenever the user saves any edit to the backup object/DLE.  Thus a  backup object containing an exclude list of:  

   exclude "*.doc" 

exclude "*.txt"  will show only "*.txt" in the Exclude form field, and the first exclude ("*.doc") is removed from the backup object when any edits are saved. 

• Although exclude patterns are case sensitive (except for Windows; see the next bullet item), the Amanda Enterprise catalog displays all pathnames in lowercase (even though it restores the original case). Because of this, it will appear that files you intended to exclude are being included in the backup because there were duplicate names with different cases included in the backup.  

  

Schily tar  

• The format should be just the name of the directory or file in the first level directory (of the directory being backed up) with no leading or trailing slashes, dots, etc. To exclude /path/to/top-level/foo/ from a backup of /path/to/top-level/, "foo" should be specified. This will only match "foo" in the first level directory and thus does not match "foo" anywhere else in the directory tree. Thus, there is no way to just exclude /path/to/top-level/foo/a because this is a second level down.  

• Globbing characters and simple regex, for instance [A-Z], * character can be used to match files and directories in the first level.  

  

Windows filesystems - Zmanda Windows Client  

Windows filesystems support wildcards in the exclude specification. Wildcards "*" (match one or more character) and "?" (match exactly one character) are supported.  The pathname in excludes specification can be absolute or relative to the DLE directory. For example: We are backing up C:\Data directory. We would like to exclude all files with *.jpg extension. The exclude specification should be "*.jpg". You can specify the directory name with the exclude pattern. For example: To exclude *.exe from C:\Data\Test and C:\Data is being backed up, specify "Test\\*.exe", To exclude a folder, specify "C:\users\all users\".  

The list of patterns in the exclude specification for Windows file systems should be separated by a space character.  

When specifying exclude patterns for Windows clients, pathnames are case-insensitive. The ./ or .\characters do not work.  

  

Reasons to Exclude Files  

 

There are a number of different reasons to exclude files from a backup set. 

 

1. You do not want the excluded data to be part of any backup set. Such data tends to be quite small in quantity and does not save much.  
2. You want to exclude the data from Compression or Encryption options, saving CPU cycles, network bandwidth, and total time to run backups.  
3. Lastly, if the data on the host is organized in such a manner that implementing multiple top-level directories in different disk lists makes no sense, exclude lists may allow users to organize the backup with the possibility of faster restore for some files. The files would be excluded from the general backup and kept in separate backup object/DLE(s).  

  

Comments:  

This is a free text field to include any comments on the Source.  

Advanced Switch:  

Disabling this switch hides advanced options. Is enabled by default.  

  

Advanced Options  

This section describes all of the possible Advanced Options that may be displayed for any of the object types. Some of these options may not apply (or be displayed in the dialog) depending on the type of backup being configured.  

Backup Estimate  

Selects the method used for estimating the backup window. You can choose from a number of options that balance the requirements of accuracy vs. speed. The "fastest" method can be accurate enough if the backup source remains relatively constant in size; the "Always Accurate" option may be too slow given the backup window, or may not be available from the given backup client. Options not supported for a given backup client are grayed out so that you cannot select them.  

Backup Strategy  

You can specify the backup level restrictions for the backup object such as perform only full backups, perform only incremental backup. You can also skip backup of backup object using this field.  

Backup Staging  

Backup Staging consist of 3 options,  

·         Use Staging if possible - Selected by default  

·         Never use staging (no parallel backups if using tapes)  

·         Required (no backup is space if insufficient space)  

Some backup objects such as NDMP backups do not support Staging Area. They can be disabled on a backup object basis using this field.  

  

Amanda Backup Client Application  

Amanda can perform backup of objects using multiple applications. There is a default application for each backup object. You can use different backup application for Linux, Solaris file systems and Oracle.  

The override field should be used only when Zmanda Support instructs you to do so. This field is usually used for custom applications.  

  

Extended Attributes  

This is the default for Windows and Mac OS X filesystems.   

This is an optional feature for Linux, UNIX, and Solaris filesystems. Enabling this option in these cases selects a different archive program used for backing up the given object type. When this option is enabled, Amanda Enterprise uses (and requires) Schily tar instead of GNU tar as the archive program. Schily tar is required on the Amanda client and is not installed by default on Linux.  Schily tar package is available for download from Zmanda Network.   

If the backup object is a Solaris ZFS file system, Extended Attributes refer to ZFS Access Control Lists (ACLS); see Solaris Client for details.  

Please be aware that exclusions are not supported when selecting Extended Attributes for Linux, UNIX, and Solaris file systems.  

  

Backup Sets:  

This field lists all the Backup Sets available and the Source can be bound to multiple backup sets.  

  

Tags:  

All the available tags will be listed to select relevant tag/s for a particular Source.  

  

Save Button:  

Save button will be enabled upon filling up all the mandatory fields with valid inputs and clicking it creates a Source with all the particulars provided.  

  

Cancel Button:  

Clicking on the Cancel button takes the user back to the Sources page. Particulars entered in all the fields will be lost upon cancellation.