10
Node Maintenance

A regular schedule of node maintenance is the best protection against unscheduled down time and loss of data. Following the procedures outlined below will minimize problems and ensure that your calendar server runs smoothly and without interruption.

This chapter outlines the following tasks:

• Maintenance procedures
• Back up and restore

Maintenance procedures

Daily monitoring procedures

The following system monitoring procedures should be performed on a daily basis:

• Check that all relevant daemons/services are operational.

  The unistatus utility displays the current status of the calendar server. For full information on use and syntax, see the calendar server Reference Manual, Appendix C, "Utilities."

• Check that ample space is left in the /users/unison directory or file system. For more information on calculating the storage requirements for your node, see Appendix A, "Disk Space and Memory".
• Verify that the previous night's backup has run.
• Search for unusual entries in the log files in the /users/unison/log directory. This task can be automated by "grep"ing/searching the log files for specific errors, and e-mailing the results to the calendar server administrator.
• Check for recent writes to the /users/unison/log/dbv.log. This file is created only if there is a problem and should be manually removed once the problem is resolved. If the file is present and is not empty, you might analyze the contents and use the unidbfix utility (contact your support provider for the most current version), or consult your support provider for further assistance.

NT

• Windows NT's Performance Monitor tool can be used to chart or log the performance and activity of the calendar services. Windows NT's Event Viewer records any problems encountered running the application.

Daily maintenance procedures

A nightly backup of the calendar database (/users/unison/db) and configuration files (/users/unison/misc) is your best protection against database corruption that may occur as a result of a power failure or disk crashes. While database corruption is rare, even under the aforementioned conditions, nightly backups serve as a safeguard in the event that your database cannot be restored. For more information, see "Back up and restore".

Monthly maintenance procedures

The following system maintenance procedures should be done after hours on a monthly basis:

• Archive the log files. Remember to shut down the server before archiving the log files, and to restart the server once the task is completed.
• If you are using a directory server, run unidssync as required to ensure that the information in the node(s) is synchronized with that in the directory server. For full information on use and syntax, see the calendar server Reference Manual, Appendix C, "Utilities."
• To improve performance and minimize disk space requirements, the unirmold utility should be run monthly to remove all events and tasks older than 12-18 months. For full information on use and syntax, see the calendar server Reference Manual, Appendix C, "Utilities."

Other maintenance procedures

• Verify the consistency of the server database(s) using the unidbfix utility. Contact your support provider for the most current version of this comprehensive database maintenance and repair utility.
• The unidbfix utility should be run in check mode once a week with the calendar server running, and in fix mode once a month with the calendar server down. If the weekly check discovers an error, it should be corrected immediately using unidbfix in fix mode; if the weekly check produces a warning, maintenance can be delayed until the monthly fix.
• If you are using a directory server, run unidsdiff to detect and resolve any discrepancies in the mapping between users and resources in the directory server with those in the calendar server node. For full information on use and syntax, see the calendar server Reference Manual, Appendix C, "Utilities." You should perform this synchronization procedure every 2-4 weeks or as required when making a batch of changes to the calendar node, particularly when deleting users. You may also synchronize your calendar and directory servers through the Admin GUI.
• To view elapsed time and CPU statistics for each client connection, set [ENG] stats = TRUE in unison.ini. When a client connection is closed, stats results are appended to the /users/unison/log/stats.log file. After analysis, set [ENG] stats = FALSE to disable logging, as the file grows quickly.

Back up and restore

To minimize the impact on your users, back up your calendar server only during periods of low user activity. If you use an external directory server, back up your directory server concurrently with your calendar server to minimize inconsisencies should it become necessary to restore a backup.

You have three options for backing up your calendar server:

• the unidbbackup utility
• stopping the calendar server and running the uniarch utility
• stopping the calendar server and copying or zipping the database files directly

The unidbbackup utility is recommended, as it provides on-line or `hot' backups, important in the context of installations requiring 24-hour access to their calendar data. An on-line backup can not be achieved by simply copying the database files while the server is still running, as the files on disk are not necessarily an accurate reflection of the state of the database at any given time. If you choose to copy the database files directly, you must stop your server to allow all database contents to be written to the disk first.

While unidbbackup is running:

• clients connected when the backup is initiated are permitted to log off
• clients not connected when the backup is initiated are not permitted to log on
• logged-on users may view but not modify their agenda
• if more than one node exists on a host, each node is locked and backed up in succession

unidbrestore is the complementary utility used for database restoration. For full information on the use and syntax of the unidbbackup and unidbrestore utilities, see the calendar server Reference Manual, Appendix C, "Utilities."

To back up a calendar host:

• Execute the unidbbackup utility through the command line or using the Database Backup dialog box of the Admin GUI. A backup will be made of all database and configuration files on your calendar server. If more than one node exists on the host, unidbbackup will back up each in turn.

To restore a calendar host:

1. Shut down the server.
2. Run unidbrestore to restore the backup. Your calendar database and configuration files will be restored to the /users/unison directory on the host.

   Warning: This operation restores only the database and configuration files. Calendar data stored in a directory server must be restored separately. If you have any reason to expect that inconsistencies may exist between the data in the calendar server and that in the directory server, use the unidsdiff and unidssync utilities to identify and resolve all discrepancies. For full information on use and syntax, see the calendar server Reference Manual, Appendix C, "Utilities."

Archived backups should be managed to ensure full data recovery capabilities without sacrificing large amounts of disk space. Remove backups that are no longer needed.