Backup/Recovery Services | Unix Services

I. Overview

UNIX backup services are managed by the Information Resource Department for the entire University. There are two backup servers: one for the EC building, and one for the remainder of the University. Each server has a tape stacker (Digital Linear Tape - DLT) attached onto which clients' filesystems are written.

II. Policy

The main objective of IR backups is to provide thorough, reliable disaster recovery of critical files, i.e., home directories, email, software development areas. Email is kept for only two weeks, while all other files are maintained for 5 weeks before the tapes are overwritten.

Users are STRONGLY urged to prepare individual backup strategies and not rely on IR backups as the sole fallback position. Whenever possible, files should be duplicated on a separate filesystem at regular intervals. Projects which maintain large datasets should purchase equipment and media to provide for their own backups. IR is available for consultation on these matters.

III. Schedule

The basic schedule is 1 full backup every other week with incremental backups in between. Backups have a substantial impact on the campus network and are therefore scheduled at night. Full backups generally commence about 00:30, and run every night with clients staggered throughout the week.

IV. User Recovery (Legato Networker/Solstice Backup)

Networker maintains an online index of each client. This index is updated with filesystem information whenever a backup is completed, and users may browse the indexes to find files/directories for recovery. Online indexes are preserved for 30 days - for recovery of files older than that, contact Assist (x2911, assist@utdallas.edu) for assistance. There are two commands which can be used by users to recover files:

newrecover

an X-windows GUI with good online help/instructions

path: /usr/local/bin/nwrecover (/usr/soft/this/sol2/nsr-5.1/usr/bin/nwrecover)

recover

a command line version with help available by typing '?'

path: /usr/local/bin/recover (/usr/soft/this/sol2/nsr-5.1/usr/bin/recover)

The steps for recovering data from Networker via the recover command are:

From the client machine (where the files were located originally), enter the command "recover". This begins a session between the client and server, and the appropriate online index is activated. The usual shell prompt is replaced with "recover".
Use the Unix commands "ls", "ll", and "cd" to navigate through the image of the filesystem.
Mark files/directories to be recovered with "add ", then initiate the recovery with "recover".
apache% recover
recover> ls
tmp myfile.dat subdir junk
recover> add subdir
adding subdir to recovery list
recover> recover
recovering...

Any difficulties with this system should be reported through the x2911 assist system.



I. Overview UNIX backup services are managed by the Information Resource Department for the entire University. There are two backup servers: one for the EC building, and one for the remainder of the University. Each server has a tape stacker (Digital Linear Tape - DLT) attached onto which clients' filesystems are written. II. Policy The main objective of IR backups is to provide thorough, reliable disaster recovery of critical files, i.e., home directories, email, software development areas. Email is kept for only two weeks, while all other files are maintained for 5 weeks before the tapes are overwritten. Users are STRONGLY urged to prepare individual backup strategies and not rely on IR backups as the sole fallback position. Whenever possible, files should be duplicated on a separate filesystem at regular intervals. Projects which maintain large datasets should purchase equipment and media to provide for their own backups. IR is available for consultation on these matters. III. Schedule The basic schedule is 1 full backup every other week with incremental backups in between. Backups have a substantial impact on the campus network and are therefore scheduled at night. Full backups generally commence about 00:30, and run every night with clients staggered throughout the week. IV. User Recovery (Legato Networker/Solstice Backup) Networker maintains an online index of each client. This index is updated with filesystem information whenever a backup is completed, and users may browse the indexes to find files/directories for recovery. Online indexes are preserved for 30 days - for recovery of files older than that, contact Assist (x2911, assist@utdallas.edu) for assistance. There are two commands which can be used by users to recover files: newrecover an X-windows GUI with good online help/instructions path: /usr/local/bin/nwrecover (/usr/soft/this/sol2/nsr-5.1/usr/bin/nwrecover) recover a command line version with help available by typing '?' path: /usr/local/bin/recover (/usr/soft/this/sol2/nsr-5.1/usr/bin/recover) The steps for recovering data from Networker via the recover command are: From the client machine (where the files were located originally), enter the command "recover". This begins a session between the client and server, and the appropriate online index is activated. The usual shell prompt is replaced with "recover". Use the Unix commands "ls", "ll", and "cd" to navigate through the image of the filesystem. Mark files/directories to be recovered with "add ", then initiate the recovery with "recover". apache% recover recover> ls tmp myfile.dat subdir junk recover> add subdir adding subdir to recovery list recover> recover recovering... Any difficulties with this system should be reported through the x2911 assist system.	Information Resources Unix Services Front Campus Unix Systems Unix Introduction Services Postmaster Services Web Services Support Backup/Recovery Services Email Services Spam Services
UTD Home · University Search · A-Z Index · Campus Map	Contact CATS