This chapter concerns performing backup and restores in a NetWare 6 environment. The objectives important to this chapter are found on page 12-1:
Concepts:
Set up SMS for SBCON and NWBACK32Novell Storage Management Services (SMS) is the category of services that include backup and restores. By establishing this type of service on the network, Novell enables an administrator to use Novell's own backup utilities, or install and use a preferred utility, as long as it is SMS compliant. It is possible to make backup copies for eDirectory, the network file system, and workstation file systems. The text begins the discussion of doing this by describing the components of SMS. The text explains that the first SMS component is a Storage Management Engine, which has three parts:
The second SMS component is the Target Service Agent. Some terminology is required to understand what this is. If we are performing a backup:
A Target Service Agent is software for a specific kind of target. NetWare 6 provides several TSAs for specific targets:
This is a rather unusual use of the word target, but it makes sense in that a target holds the data you are backing up or restoring. It would be too simple if the software already described was sufficient, so the third component of SMS is introduced: the Storage Management Data Requester (SMDR) fits between a TSA the the SME. Finally, the fourth and fifth components are storage device interfaces and drivers, to enable the SME to save or read data that is saved. Data SetsThe concept of data sets applies to eDirectory and to file systems. Data sets and subsets can be defined to make specific backups of data. In general, a data set can be a parent or a child. Parent data sets contain other data elements, child data sets do not. This is different from the usual definition of parent and child. Under this definition, a child is only a collection of files or eDirectory leaves. Any object that contains other data objects is a parent, regardless of what we would normally refer to as its child relationship to a higher object. For example, the SYS: volume is a parent data set, and so is any directory in it. Those directories are not considered children of SYS: under this odd definition. Backups can use include and exclude options. Use the include option to back up only a small part of the data system. Mark the part you want to be included. Use the exclude option to back up everything except the part you mark to be excluded. You can combine include and exclude instructions to customize how much information will be backed up. Enabling SMS to Use SBCON and NWBACK32The text gives a three part procedure to use the backup utilities with SMS. Note that SMS is not loaded until the third part.
To use SBCON, load it at a server console. When done, exit the program interface, then unload SMS related NLMs with the command SMSSTOP. The text notes that you could get an error message with this command, that an NLM you are trying to unload is in use. If you get a message like this, and tell the server to unload the NLM, you could abend the server. NWBACK32 is used from a workstation. To use NWBACK32, you should log in to the server that runs SMS. In its Public directory, find and run the NWBACK32.EXE program. Use and exit the program as you would any other Windows program. At the end of the session, issue the SMSSTOP command, making sure not to abend the server, as explained above. Guidelines for Backup UtilitiesWe are given guidelines for using NetWare Backup/Restore. To summarize them:
When you run a backup or restore, log and error files are created in the SYS:\SYSTEM\TSA directory. These files can be accessed from SBCON and from NWBACK32. Log and error files from a backup procedure are stored under SYS:\SYSTEM\TSA\LOG, unless you tell the server to store them elsewhere. Log and error files from a restore procedure are stored under SYS:\SYSTEM\TSA\RESTORE. This location cannot be changed. Back Up Data with SBCON and NWBACK32The next topic repeats some of the information above, and adds some new details. Using SBCON to perform a backup:
NWBACK32 is run from a workstation:
Verifying a BackupYou can verify that a completed backup job has no errors from either the SBCON or NWBACK32 interface. There are many steps to do so, and you should be aware that the real test of a backup is to perform a restore. Backup StrategiesThree backup strategies, or schedules, are explained. You should
know all three. First some terms:
This needs more explanation. Assume we use a tape drive to make backups. In a Full backup strategy, the entire target is backed up to tape every time we make a backup tape. This strategy consumes the most time and the most tapes to carry out a backup. To restore, we simply restore the most recent tape(s). This is the least time consuming strategy for restoring, but the most time consuming for creating backups. The second method, Incremental backup, means that we start with a Full backup of the target, and then each successive backup tape we create only backs up the elements that are new or changed since the last backup was created. This means that successive backups will not always be the same length. Therefore, this is the least time consuming backup, but the most time consuming restore. To restore, we must first restore the last Full backup made, and then restore EVERY tape made since then, to ensure getting all changes. (Incremental backups can only be done for file systems, not for eDirectory.) The last strategy, Differential backup, also starts with a Full backup tape. Then each successive tape made will contain all the files changed since the last Full backup was made. This means that we will have to restore only one or two tapes in a restore operation. If the last tape made was a Full tape, we restore only that one. If the last tape made was a Differential tape, we restore the last Full tape, then the last Differential tape. (Differential backups can only be done for file systems, not for eDirectory.) In both Incremental and Differential backup strategies, you will typically use a rotation schedule. For example, you could have a one week cycle. Once a week, you make a Full backup, then every day after that you make the other kind, Incremental or Differential. To keep them straight in your mind, remember that:
The time required to create backup tapes should be considered along with the time to restore a backup. When you consider the two concepts as two sides of the answer to a question (What method should I use?), the answer may be the most common choice: Differential. It is the best compromise in terms of backup time versus restore time. Note also, that all three methods require a full backup on a regular cycle. The recommendation is usually to run a Full backup tape weekly. A new note in this edition of the text is that trustee rights in the file system are backed up with the file system. Trustee rights saved in the file system are associated with eDirectory distinguished object names. The point seems to be that if you restore a file system, rights to files and directories will automatically be applied to eDirectory objects as long as objects exist in the tree with distinguished names that match the names saved in the file system. SYS Volume BackupA special note is made about information stored on the SYS: volume. Backing up this volume will back up several special purpose files:
These files can be backed up separately from backing up the rest of the volume. They need to be restored by name if the SYS: volume must be restored. eDirectory Backup Guidelines
Object IDsThe chapter explains in two places that eDirectory objects have object IDs, which are described as random numbers that are assigned to objects when they are created. The randomness of a number would appear to be a security feature, so the numbers are not predictable. When an object is restored from a backup, one of two things can happen.
Novell Commercial Message?The text makes a point that Novell's backup utilities allow selective backup and restore to your tree, but this feature is not found in all third party backup utilities. You are allowed to back up the entire tree, everything from a specific container to all its leaves (a branch), one container, or one leaf. This assumes that you have supervisor rights to all objects in the tree. If you do not, the utilities can scan for all objects to which you have supervisor rights in order to back them up. The information above suggests the next topic: you can assign various users the supervisor right to various containers, which will allow them to make backups of those containers and the objects in them. To make this easier for these users (Novell calls them "backup administrators"), you create a file called SYS:\SYSTEM\TSA\TSANDS.CFG. Each line in this file will hold the typeful distinguished name of a container. Put a line in the file for each container that you assign to a backup administrator. This will make it easier for the backup administrators to back up only the branches they are responsible for. Restore Data with SBCON and NWBACK32Some of the same information needed when making a backup is needed when
performing a restore. The chart in your chapter is merely a repeat of
the one from the backup section. It should probably look more like this:
The steps to perform a restore using SBCON and NWBACK32 are approximately the same as those for performing a backup. Identify eDirectory Recovery ProceduresIn the last section of the chapter, some advice is offered about restoring eDirectory. It is better to allow the replicated nature of eDirectory to make repairs than to perform restores. However, this is not always possible, so some general rules apply. If you lose eDirectory data:
If you lose a volume other than the SYS: volume, there is no direct damage to eDirectory, so a file restore is all that is needed. If you lose a SYS: volume, or your only server, that server cannot run. The server will have to be reinstalled (both NetWare and eDirectory), and then restored from backup. The basic recovery procedure is:
The procedure is a bit different if this is one server in a large tree. Procedure:
If you lose the entire tree, due to a disaster that harms all your servers:
|