Files stored on the network are backed up to tape every night. Files on tape are retained for 6 months, after which time the tapes are reused, and files are overwritten. There are some exceptions, and they are noted below in greater detail.
- Purpose
- General Information
- Accidental Deletions, Overwrites, and Corruption
- Schedules
- Retention
- Recoveries
- File Versions and Recoveries
- Users Responsibilities for Recovering Files
- Time to Recover a File
- PC Workstations
- Potential Recovery Problems
Purpose:
The primary purpose for the backup system is to provide for disaster recovery of key network servers and services (applications, databases, web pages, and servers hosting user's home and group directories). Additionally, the backup system provides recovery of any file that has existed on a network server long enough to be caught in a regularly scheduled backup of that server (usually a minimum of 24 hours). The backup system is not an archival system for storing information off-line for indefinite periods of time.
General Information:
Backups are scheduled to run each night. Recovers can be done at anytime during the day when the backup system is idle. The earliest point in time that the backup system can recover a given file is from the most recent successful backup of that file. Several rare variables may prevent a file from being backed up successfully. These include, but are not limited to, network outages, file corruption, or the file being in an "open" state when it's backed up. Backups fall into one of three categories: 1) Baseline - A full backup of every file on a give network server. 2) Level - A consolidation backup of any file that changed, or was created, since the last level or baseline backup. 3) Incremental - A backup of any file that changed, or was created, since the last backup performed.
Accidental Deletions, File Corruption, and Accidental Overwrites of Files:
A file that has been accidentally deleted can only be recovered from the backup system if that file existed on the network at the time of the last backup. In this event, the file will be recovered in the state it was in at the time of the last backup. Any new changes to the file are not recoverable by the backup system. However, all is not lost. In many cases, a system administrator can use other tools (like Emergency Undelete services) to recover an accidental deletion. This will not be able to restore different versions of the file though (i.e. the state of the file between various "saves" you've done throughout the day). Also, if the file is overwritten, it will not be available to the Undelete service in the state it was in prior to being overwritten. In the case of corruption or overwrites, you will need to have the file restored from tape.
Schedules:
Backups are performed each night, usually performing an incremental backup, backing up only the files that changed, or have been created, since the last backup. Baseline and level backups are staggered to run throughout the month to reduce backup traffic on the network. Backups for servers providing specific services may follow different schedules, more or less frequently depending on need, to ensure recoverability of that service (see table below).
Data Retention:
Retention is the period of time that files can be recovered from tape, after which the tape is subject to being recycled (reused) and old backup data is over-written with new backup data. Backup tapes are kept available for recoveries for a minimum period of 6 months before being recycled. After that, tapes are recycled (information is overwritten) as necessary to keep the backup system running. This means that, depending on the type and location of the file, files may be recovered at any time for a period of 1 to 6 months, after it was last seen on the network and backed up by the backup system. The following table describes the types of files being backed up, when they're backed up, and how long they're retained.
Retention and Schedules
Files Backed Up | Baseline | Level | Incremental | Retention Period |
---|---|---|---|---|
Home Directories | Monthly | Weekly | Daily | 6 months |
Group Directories | Monthly | Weekly | Daily | 6 months |
Web Directories | Weekly | n/a | Daily | 1 month |
Databases | Nightly | n/a | n/a | 1 month |
Applications | Quarterly | n/a | n/a | 3 months |
System Files | Nightly | n/a | n/a | 2 weeks |
Common/Share space | Monthly | Weekly | Daily | 6 months |
Temporary Directories | Weekly | n/a | Daily | 1 month |
Tapes are stored in both on-site and off-site locations for security.
Recoveries:
Any file stored on a network server should be "recoverable" from tape after it has entered the backup cycle, and as long as the tape has not exceeded its retention period. This requires that the file exist for a minimum of 24 hours in order to make it into the backup cycle. Accidental deletion, corruption, or overwrites, of newly created files that have not made it into the backup cycle cannot be recovered from system backups.
Recovery of a file that has not entered the backup cycle may be possible through third party system administrator tools such as Emergency Undelete. Support staff cannot guarantee these tools will be successful, especially in the case of overwrites.
A request for a recover must be made to support personnel (helpdesk). It is not possible for individuals to recover their own files for security reasons (special circumstances and exceptions exist, and are noted below).
File Versions and Recoveries:
Because of the method in which files are backed up, multiple versions of a file may exist in the backup system. When a recover is requested, the most recent version of a file that can be successfully recovered will be restored to the network server. This may result in a partial loss of work depending on the state of the file during the backup (if it was open) and the amount of work done on the file between the last successful backup and the point of failure.
The backup system cannot recover modifications to a file made between the last successful backup and the point of failure.
Point of failure refers to the point in time in which the file becomes deleted, corrupted, lost, etc. for whatever reason.
User Responsibility for File Recovery:
Users must provide support personnel with three pieces of information in order for any file to be recovered:
- The full directory path to where the file was located
- The name (spelled correctly) of the file to be recovered
- When the file was lost, deleted, overwritten, or corrupted
Files cannot be recovered until support personnel have the above information.
Time to Recover a File:
Every attempt will be made to recover a file that has been backed up, and that is recoverable, within a reasonable period of time. This can take from several hours to several days. A file is considered "recoverable" as long as it exists on a tape that has not been recycled. Instant recovery is not possible.
PC Workstations:
Individual PC Workstations are not included in the network backup system. Users with sensitive files, that require backup, should keep those files on the network.
However, on a cost-recovery basis, users may elect to have their individual PC's participate in the network backup system. This will require special software and licensing, and may require additional approved hardware depending on individual needs. Users will ultimately be responsible for the backup of their PC's, but also have the ability to perform their own recoveries.
Recovery Problems:
Rare recovery problems do exist that are beyond support staff's ability to control. In some cases, files may not recover properly. This can be due to many factors, and may result in the file being unrecoverable or may affect the time required to recover a file. Factors affecting recovers:
- Incomplete backup
- File was "open" or "in use" during the backup
- Errors writing to the media during the backup
- Corrupt file prior to backup
- Network errors during recovery
- Disk errors during recovery