|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Navigation Bar
|
TSM Backup HintsScheduling - controlling the start time
Scheduling - controlling the start timeEver been in the situation where you have a tight backup window (and if you haven't, how do you get a job like that?) You schedule 8 backups to start at midnight, but some of them don't start until 01:00? Frustrating or what? The problem is that by default, TSM tries to spread backups out in a schedule, unless you tell it not to. Issue the command QUERY STATUS on your TSM server, and look for the parameter Schedule Randomisation Percentage. By default, this will be set to 25, which means TSM will spread the start times of all the backups in a schedule, over the first 25% of the schedule window. If you want all your backups to fire in right at the start of the schedule, then change this parameter to 0. If you get a problem with the schedule, the error will be either 'failing' or being 'missed'. 'Failed' means that the scheduled event began, but encountered an error condition that prevented it from completing successfully. 'Missed' means that the scheduled event never began. For example, if the scheduler is not running on the client, then of course it can not execute its scheduled events; nor can it report that it did not run. On the client machines that missed their schedule, verify that the scheduler service is started. Check the dsmsched.log and dsmerror.log to verify that the scheduler was able to successfully contact the server and query its next scheduled event. Scheduling - choosing specific weekdaysIf you define your client schedules using the GUI interface, you can chose to run a backup every day, weekdays only, or on a specific day of the week. You can only chose one of these options. If you are running TSM version 3 and you use a command line, you can be more selective. Say you want to define a schedule to run every Monday, Wednesday and Friday. Use a command like DEF SCH domain_name schedule_name T=C ACT=I STARTT=22:00:00 DUR=2 DAY=Monday,Wednesday,Friday Compression space errorsPrior to a client sending a file, the space (same as allocated on client) is allocated in the TSM server's disk storage pool. If caching is active in the disk storage pool, and files need to be removed to make space, they are. But if the file grows in compression (client has COMPRESSIon=Yes and COMPRESSAlways=Yes), the cleared space is insufficient to contain the incoming data.
This commonly happens where client compression is turned on, and client has large or many compressed files: TSM is fooled as compression increases the size of an already-compressed file.
How to determine when TSM will expire a backupThe key to understanding TSM backup retention is to understand the difference between 'active' and 'inactive' backups. The 'active' backup is the most recent backup while all the older backups are 'inactive backups'. However, once a file is deleted from the client, it becomes inactive after the next incremental backup. The active backup of a file is never deleted, as it is needed to recreate a disk
So the pecking order goes like
For example, you have RETEXTRA=31 and VEREXISTS=31. If you create 31 versions in the same day, and then (still on the same day) you create version 32, version 1 will expire, regardless of RETEXTRA, because the VEREXISTS criterion has been exceeded. Likewise, if you create version 1 today, then create version 2 a week later, then never create another version after that, then version 1 will expire 31 days after the creation of version 2, since the RETEXTRA criterion has been exceeded. If you need to GUARANTEE data retention for 31 days you would need to code the parameters as below, but be aware that you could end up keeping a lot of backups.
Incremental by Date backups; faster but less secureYou would use this type of backup if your backup window is not long enough during the week, but you have plenty of time at the weekend. An incremental by date backup uses the last updated timestamp on a file to decide if it needs to be backed up or not. The problems is that this field is not one hundred percent reliable on open Systems data as some applications can update data without changing the last update field. A normal TSM backup will compare the attributes of every file with the current active backup, and if they do not match, will take a new backup. Incremental by date simply looks at the last modification date, so it is much faster, and uses less memory. The downside is that is might not backup every changed file. Also, it will not expire backup versions of files that are deleted from the workstation, it will not rebind backup versions to a new management class if the management class has changed and it ignores the copy group frequency attribute of management classes To run an incremental by date backup, you add the parameter '-incrbydate' in the 'OPTIONS' box in the 'Define Client Schedule' GUI window. If you use incremental by date, then to ensure you do get a full client backup, and correctly manage file expiration and management class changes, you should plan to take at least one full incremental backup every week. An NT drive is refusing to allow a backupThis probably a permissions problem. You need to check what account the scheduler service uses to log in, then make sure the drive has permissions for that account. If you are working with a Microsoft cluster, then for each logical node on the cluster you need to include CLUSTERnode Yes in the cluster.opt file. You also need to make sure that the registry replication is set up in the scheduler service, so the generated password is replicated to both nodes in the cluster. You can check this by issuing a command from either cluster node, and seeing if it connects to the TSM server correctly.Incremental backup of multiple directories using the objects fieldTo perform an incremental backup for several directories, code them in the objects field in the schedule like this. OBJECTS='/path1/ /path2/ /etc/' The terminating slash at the end of the directory name backs up the files within the directories, and they need to be separated by spaces. If you wish to back up subdirectories within these directories, then you also need OPTIONS=-SUBDIR=YES Querying BackupsetsThis will give you the list of backupset volumes from volhistory:
select volume_name -
from volhistory -
where type='BACKUPSET'
This will give you the NUMBER of backupset volumes from volhistory:
select count(volume_name) -
from volhistory -
where type='BACKUPSET'
This will give you a list of the backupset volumes that are NOT checked in:
select volume_name -
from volhistory -
where type='BACKUPSET' and volume_name not in -
(select volume_name from libvolumes)
This will give you the NUMBER of backupset volumes that are NOT checked in:
select count(volume_name) -
from volhistory -
where type='BACKUPSET' and volume_name not in -
(select volume_name from libvolumes)
Adding a new Management ClassManagement Classes are used to determine how long backups are kept for, and where the backup data goes. To understand Management Classes you need to know how they fit in.
Each policy domain will contain one or more Policy Sets. Only one Policy set will be active in each Policy Domain and it is actually called 'ACTIVE'. You cannot update the active policy set. The other Policy Sets are used for version control.
So suppose you want to add a new Management Class called MCDB2_TAPE that will allow DB2 backups to go LAN free direct to tape, and you want that to be the default Management Class. You already have an exisiting DB2 Policy Domain called PSDB2 with Existing Management Classes. The first step is to make a copy of the Active Policy Set called PSDB2_MAY2010, to make sure you keep the existing classes. (create a working copy of the policy set) COP PO PDDB2 ACTIVE PSDB2_MAY2010 (Define the new management class to that set) DEF MG PDDB2 PSDB2_MAY2010 MCDB2_TAPE DESC="DB2 TAPE Management Class" (Define a backup copygroup. The retention parms leave retention control with DB2) DEF CO PDDB2 PSDB2_MAY2010 MCDB2_TAPE TYPE=BACKUP DEST="PRIMARY_TAPE_POOL" - VERE=1 VERD=0 RETE=0 RETO=0 (Assign the new management class as the default) AS DEFMG PDDB2 PSDB2_MAY2010 MCDB2_TAPE (Validate the changed Policy Set) VAL PDDB2 PSDB2_MAY2010 MCDB2_TAPE (Activate the new Policy Set) ACT PDDB2 PSDB2_MAY2010 MCDB2_TAPE Of course there is a lot more to LAN free than just defining a Management Class, but LAN free will not work unless the target files are directed straight to tape. Using Different Management ClassesOne way to bind a set of backups to a different management class is to add an include statement into the client options file with a statement like INCLUDE "C:\Program Files\Microsoft SQL Server\MSSQL\Backup\...\*" MCSQLBK means bind all files and files in subdirectories in the Backup directory to special management class MCSQLBK. If you add this statement, you will bind all previous backups of these files to the new management class. The '\...\' means scan subdirectories The rebind happens next time a backup runs. This will work for every backup version of the file, not just for the active one. The file must
be examined again to get a new backup class, so you cannot change management classes for files that have been deleted from the client.
part of the statement means include subdirectories.
back to top Selecting and Excluding DrivesYou can select and exclude drives from scheduled backups by placing an entry in the DSM.OPT file on the client, with a line which looks like this for a Netware client DOMain DATA: APPLICS: SYS: or for a windows client DOMAIN c: d: The problem with these approaches is that you need to remember to update the dsm.opt file if you add new drives DOMAIN ALL-LOCAL will backup everything. A good variant, if you never wanted to backup your SYS: drives for example is DOMAIN ALL-LOCAL -SYS: This means that all drives are backed up except the SYS: drive, and you do not need to change the dsm.opt file as drives are added or removed. Specific selection criteria for the Windows System Object, or the Netware eDirectory (NDS) are given below. However, this approach will only work if you are backing up at domain level, which is the normal approach with a scheduled backup. If you specifically select a volume that is excluded in the dsm.opt file, then it will be backed up. dsmc incremental sys:\ -subdir=yes will backup the entire SYS: volume, even though it is excluded in the domain statement. This is not necessarily a bad thing, it means that you can exclude SYS: from scheduled backups, but when you really want a backup, you can do it manually. This approach will also allow you to backup selected files or directories from an excluded disk dsmc selective sys:\tivoli\tsm\* -subdir=yes If you absolutely never want to let anyone backup the SYS: volume in any circumstances, then use EXCLUDE statements as these will always apply. To exclude an entire disk you need two commands, one to exclude the directories and one to exclude the files in the root folder, like this (Technically, this does not quite exclude the whole disk as the root of the volume itself will be backed up once). exclude sys:\* exclude.dir sys:\* or exclude c:\* exclude.dir c:\* So now if you try to backup with the command dsmc selective sys:\tivoli\tsm\* -subdir=yes nothing will happen, because those files are always excluded. back to top Including and Excluding dataThere are two types of INCLUDE and EXCLUDE commands, INCLUDE, EXCLUDE, INCLUDE.DIR and EXCLUDE.DIR. The first two commands will include or exclude files, the second two commands include or exclude directories. Syntax is INCLUDE c:\file\space\name\..\* EXCLUDE.DIR c:\file\space\name\temp Note that the directory exclude does not end with a '\'. The inclexcl list is normally processed from the bottom up but EXCLUDE.DIR is processed before the regular INCLUDE or EXCLUDE statements so the position of EXCLUDE.DIR does not matter. However it is best to put the DIR processing statements at the bottom of the list to make it more obvious how the processing works. If the data path you are describing includes spaces, you must include the full statement in quotes, i.e. INCLUDE"C:\Program Files\Microsoft SQL Server\MSSQL\Backup\...\*" The \...\ means include subdirectories. If you just coded Server\MSSQL\Backup\* then the include would only apply to files within the backup directory, and not any files in subdirectories.
EXCLUDE \filespec\*.* will only exclude files which have a period in the name. If you want to exclude all files in a path, you need EXCLUDE \filespec\* Recent versions of the TSM client software provide an exclude.fs statement that will exclude an entire file system. It is more efficient than using a plain exclude statement. An exclude.fs suppresses any examination of the directories within the file system. An exclude covering an entire file system does not. The TSM client will still read every file name in every directory within the file system. It will check each file name against the include and exclude statements. It will then decide not to back up that file (assuming the exclude for the entire file system is in the right place in the sequence of include and exclude statements). Another recent addition is the query inclexcl command, which lets you check your syntax. This was introduced in TSM 4.1 It was still possible to check syntax on older releases, using the unsupported show inclexcl command. Another variant is to use ranges within EXCLUDE commands, for example, say you wanted to backup a large disk with three backup streams, but you did not want to have to change include statements when adding new directories (it would be more likely that you would not know when new directories were added, so they would be missed). You could define three clients, each with its own dsm.opt file, with exclude statements as shown below. You would need to ensure that the ranges covered every possible directory name. Content of dsm_a.opt NODENAME Servername_A EXCLUDE.DIR \\Servername\diskname\[K-O]* EXCLUDE.DIR \\Servername\diskname\[P-Z0-9$]* Content of dsm_b.opt NODENAME Servername_B EXCLUDE.DIR \\Servername\diskname\[A-J]* EXCLUDE.DIR \\Servername\diskname\[P-Z0-9$]* Content of dsm_c.opt NODENAME Servername_C EXCLUDE.DIR \\Servername\diskname\[A-J]* EXCLUDE.DIR \\Servername\diskname\[K-O]* Be aware that if you are running TSM version 5.4.0.0 or 5.4.0.1 and change your include-exclude options using the client setup wizard or the Windows backup-archive GUI client preferences editor the order in which INCLUDE and EXCLUDE options occur in the options file will be reversed. You will almost certainly end up missing files that should be backed up. Your best option is to install version 5.4.0.2 TSM fails with out of memory errors on the backup client.If you are having problems with TSM running out of client memory then you have a number of options to fix it. TSM uses about 300 bytes in memory for every object it backs up in a file system. This means that a 2GB virtual memory will support about 7 million objects. Note that every operating system will have a maximum virtual memory limit which places an upper limit on the number of objects that can be backed using default parameters. One solution is to specify memoryefficientbackup yes in the dsm.opt file. This changes the processing so now TSM only needs enough memory to process all the objects in one directory, not the whole file system. However this option does have a performance overhead. Another option is to use incrbydate, described elsewhere on this page. Other options include using Windows journalling backups, TSM image backups or on a UNIX system, define multiple virtual mount points within one file system. Each mount point would be backed up independently. Selective backups of Windows directories with embedded spacesIt is possible to select a number of directories from a Windows command line interface by listing them as in the example command below. dsmc inc c:\dir1\* "d:\dir2\sub dir1\*" d:\dir3\ -subdir=yes This command will incrementally backup all files in directories
back to top Backing up the Windows System ObjectSee the Windows section for details of what Microsoft call the System State , and TSM calls the System Object. You backup the Windows System object by either using the command backup systemobject Or by specifying an ALL-LOCAL domain in the dsm.opt file. The ALL-LOCAL domain includes the system object. Exactly what you backup will depend on the release of Windows, and what Windows components are installed. The full list of system objects is
If you expand the System Object container in the Windows B/A GUI, you can see which system objects are active for a given client. Alternatively, you can issue the undocumented command SHOW SYSTEMOBJECT One way to check that the system object is being backed up, is to look for a 'SYSTEM OBJECT' file space on the TSM server. The main issue with system object backups is that they are always backed up every night, and the amount data can be about 2GB. You may want to restrict the number of backups held, by binding the system object files to a management class that keeps relatively few versions. You this with the following include statement in the dsm.opt file, or in include/exclude file if you keep these separate INCLUDE.SYSTEMOBJECT ALL yourmgmtclassname If you don't want to backup your system object every day, you can exclude the system object in the dsm.opt file by prefixing it with a '-' (minus) sign. (If you are running TSM 5.1 or higher). DOMAIN ALL-LOCAL -SYSTEMOBJECT Then once a week, you would run a special schedule to execute the 'backup systemobject' command. You can't do this directly, you need to create a macro file that contains the command BACKUP SYSTEMOBJECT in a file called something like C:\Program Files\Tivoli\TSM\baclient\backsobj.mac and then create a schedule like DEFINE SCHEDULE STANDARD BACKUPSOBJ ACTION=MACRO OBJECTS="C:\Program Files\Tivoli\TSM\baclient\backsobj.mac" Since version 5.2.2, TSM will use the Microsoft Volume Shadowcopy Service (VSS) on Windows Server 2003 to back up all system state components as a single object. This will provide a consistent point-in-time snapshot of the system state. Changing the Management class for System ObjectsIf you want to change the management class on System Objects, you need to add a line to dsm.opt. ALL system objects must be bound to the same management class. include.systemobject ALL new_class_name If the system object won't rebind to new mgt class, try deleting the filespace, and it should rebind on the next backup. The command to do this from the server is del fi nodename "SYSTEM OBJECT" nametype=unicode Why would you want to do this? Well the system state is large, so if your default management class holds 40 backup versions say, then you will use lots of TSM backup space for the system state for every server. You can assign a management class to the system state that keeps fewer versions for less time, to save space. TSM and Windows ASRTSM can work with Windows 2003 ASR to simplify bare metal recovery, provided the TSM client is at 5.2.0 or above.
You take an ASR backup using the TSM client GUI, check the box that says 'Automated System Recovery' then click on the Backup button. Windows will then write system information to ASR state files called NTDLL.asr and SMSS.asr.
back to top Backing up the Netware NDSNovell Directory Services (NDS) is critical for the Netware server environment. You should backup the entire NDS tree periodically, and make sure that the DIRECTORY is included in the domain of at least one of your nightly, scheduled backups. TSM will not be able to recover every part of an NDS, schema extensions, partition boundary information and replica assignments are not backed up at present. Its best to keep an accurate blueprint of the NDS tree so that you can recreate this information after a disaster. TSM Version 5.1 and above uses a file space name of NDS to backup the NDS. This means that you cannot call any Netware disks NDS, or they will conflict with NDS backups. Full NDS backupTo manually backup the NDS, use the command selective nds:* -subdir=yes -volinfo Version 5.2.2.0 of the TSM client for NetWare will automatically backup NDS on a server that holds a master replica. Previous versions of the TSM client, you need to add NDS: to the domain line in your dsm.opt file, it is not included in all-local domain DOMain all-local, NDS NDS treesIf you have split your NDS tree into several branches, then its network intensive to gather data from a remote branch. It best to exclude NDS entries from remote branches, and just backup from the root, and the local branch. Say the NDS tree has three branches, US, EU and OZ, and you want to backup the local EU tree, a typical include/exclude file would then be EXCLUDE "NDS:.O=MYCOMPANY.OU=US.*" EXCLUDE "NDS:.O=MYCOMPANY.OU=OZ.*" EXCLUDE.DIR "NDS:.O=MYCOMPANY.OU=US.*" EXCLUDE.DIR "NDS:.O=MYCOMPANY.OU=OZ.*" TSM stores NDS object names as the typeful, distinguished object name including '.[Root]'. For example, the object CN=Miller.OU=Operations.O=RBS would have the following Tivoli Storage Manager name: .[Root].CN=Miller.OU=Operations.O=RBS This means that if you want to, you can exclude right down to common (leaf) objects. There are some restrictions, for example you cannot use the include and exclude.dir options together. print queuesIBM recommended that you INCLUDE the directory structure for print queues, and EXCLUDE the print files, as they cannot be backed up anyway. This means that if the server is restored, print pending documents are lost, but the queue structure is available for new requests. EXCLUDE <vol>:\QUEUES\*.QDR\* INCLUDE <vol>:\QUEUES\*.QDR\*.Q If Novell Distributed Print Services is used for printing, then Netware does not use print queues, so both the directory structure and the files can be excluded. Server Specific informationThere is some Netware information which is specific to a server, and which is not backed up by default. If you want to include it, you need to add a domain statement to the dsm.opt file DOmain serverspecificinfo The serverspecificinfo contains the following -
Also Server centric ids are unique ids assigned to each server, and are backed up server as part of the server specific information Backing up a Netware Cluster with TSMNetware Clusters have two types of disk, local disks and clustered disks. The SYS: disk will probably be local, and you may have others. The local disks are always attached to one particular server, while the clustered disks can move around the various servers in the cluster. 'Takeover Scripts' are used to make sure that the disks move cleanly between servers. TSM backups can be 'cluster aware', that is they can move with the disks as the disks move between the cluster servers. Actually, it is not the disks that move around between servers but Netware Partitions. To keep things simple, many sites set up each disk in its own Netware Partition, but your site may have several disks in each partition. When the TSM manuals refer to a 'cluster group' they really mean a Netware partition. The TSM software has to run on a physical server, but there is normally no way to decide ahead of time which physical server will be hosting a volume. The key to backing up a cluster volume is that the backup metadata must be available from whichever server is hosting that volume, so the metadata must be held on the cluster volume. The metadata includes the dsm.opt file and the password file. The schedlog, errorlog and webclient log also need to be held on the cluster volume to get continuity between messages as the volume moves between servers. Every Netware partition needs its own dsm.opt file. Backing up the Local VolumesJust use a standard TSM install on each of the physical servers. The dsm.opt file should specify CLUSTERNODE=NO (or miss it out as that is the default). With this setting, if you use a domain of ALL-LOCAL then it will not see the clustered disks. The NODENAME should be the same as the server name Backing up the Clustered VolumesEach Netware partition must be defined to TSM as a separate node, and must have a unique name that is not the same as any physical server name. As each partitions will have a virtual server name, it is easiest to use that as the node name. Allocate a TSM directory on a volume in the partition and copy a dsm.opt file into it. Assuming that your are storing the tsm info on a disk called CAV1, edit the dsm.opt file with the following settings
To set up the passwords, from your first clustered server enter the following commands:- Unload TSAFS then reload it with TSAFS /cluster=off dsmc query session -optfile=CAV1:/tsm/dsm.opt dsmc query tsa -optfile=CAV1:/tsm/dsm.opt dsmc query tsa nds -optfile=CAV1:/tsm/dsm.opt Make a copy of dsmcad in the SYS:/Tivoli/tsm/client/ba/ directory and give it a unique name for this volume, say DSMCAD_CAV1 then start the scheduler with dsmcad_CAV1 -optfile=CAV1:/tsm/dsm.opt Repeat this for every server in the cluster, and get the DSMCAD command added to the takeover scripts so the correct DSMCAD is started as a volume moves between clustered servers. Backing up Databases with TSMDatabase backups are usually a bit special as a database usually consists of a number of physical files that all need to be backed up as an entity, often with consistent time stamps. Databases also have transaction logs to ensure that the data stored in a database is consistent, even after a hardware failure. Databases usually have internal catalogs which record these files, so when you do a restore, you need to make sure that the catalogs hold the correct information too. To help with this lot, databases have a Database Management System (DBMS), which tracks physical database files, transaction logs and backups. A DBMS will usually be able to run a backup while the database is active, which effectively means no backup window is required. This means that it is not a good idea to use TSM simply to backup databases as a collection of flat files, but what you want is for TSM to use the DBMS facilities to get a good, secure backup. TSM does this by providing (chargeable) add-on modules called TDPs or Tivoli Data Protection modules. TDPs exist for DB2/UDB, Oracle, Informix and MS-SQL databases; and also Lotus Domino and MS-Exchange e-mail databases. It is also possible to interface TSM with DBMS systems for other databases using ADSMPIPE. This is described in the red paper http://www.redbooks.ibm.com/abstracts/redp3980.html Copyright © Lascon Storage Ltd. 2000 to present date. By entering and using this site, you accept the conditions and limitations of use
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
