Museum

Home

Lab Overview

Retrotechnology Articles

Online Manuals

⇒ nsr_crash(8) — DG/UX R4.11

Media Vault

Software Library

Restoration Projects

Artifacts Sought

Related Articles

nsr(8)

recover(8)

savegroup(8)

saveindex(8)

scanner(8)



NSR_CRASH(8)               Legato NetWorker 4.1.1               NSR_CRASH(8)


NAME
       NSR crash - How to recover from a disaster with NetWorker

DESCRIPTION
       NetWorker can be used to recover from all types of system and
       hardware failure that result in loss of files.

       When a NetWorker client has lost files, the recover command can be
       used to browse, select, and recover individual files, selected
       directories, or whole filesystems.  If the networker recover command
       is lost or damaged it will have to be copied either from another
       NetWorker client or from the NetWorker distribution tape.

       When recovering a large number of files onto a filesystem that was
       only partially damaged, you may not want to overwrite existing
       versions of files.  To do this, wait until recover asks for user
       input to decide how to handle recovering an existing file.  You can
       then answer N meaning ``always no'' to cause recover to avoid
       overwriting any existing files, or n if you want to protect this file
       but you want recover to ask again on other files.

       If you do want to replace the existing version of a file or set of
       files with the saved versions use the add command in recover to
       select which files should be retrieved, and answer Y or y when it
       asks if it should overwrite existing files (Y means ``always yes''
       for future overwrite cases; y means just overwrite this one file).

       For more information on using the recover command see the recover(8)
       manual page.

       If the NetWorker server daemons or commands are lost, it may be
       necessary to re-install the server from the NetWorker distribution
       tape.  Once the NetWorker server is installed and the daemons are
       running, other NetWorker server files can be recovered using the
       recover command.  When re-installing NetWorker you must be sure to
       install the /nsr directory in exactly the same place as it was
       originally installed.  The machine used to recover files may be
       different that the one used to save the files, but it must have the
       same hostname as the original machine.  This is important because
       recovery of the NetWorker on-line index requires that the index files
       have the same pathname, which includes the server's hostname, as they
       did at the time of the latest save.

       In the event that the NetWorker server's index is lost, it will be
       necessary to first recover the index from media before the recover
       command can be used to browse and recover other files.  To recover
       the NetWorker server's index use the recoverindex command.  The
       recoverindex command quickly recovers the lost on-line index for a
       NetWorker server by locating the bootstrap save set produced by the
       savegroup(8) command at the end of an automatic save.  The bootstrap
       save set tells recoverindex which save sets to extract from which
       volumes to recover the whole index.  The save set identifier and
       other information about the bootstrap save set is printed by
       savegroup at the end of each automatic save.  See the saveindex(8)



Licensed material--property of copyright holder(s)                         1





NSR_CRASH(8)               Legato NetWorker 4.1.1               NSR_CRASH(8)


       and savegroup(8) manual pages for more details.

       After recoverindex completes, the server's file index and media
       database will be fully recovered (unless recoverindex prints an error
       message to the contrary).  However, the NetWorker server's /nsr/res
       directory, which contains the resource files describing your
       NetWorker installation, require additional work on the part of the
       administrator before they will be used.  Perform the following steps
       to complete the recoverindex process (these steps will be
       incorporated into recoverindex in a future release).

       1.     Shut down your NetWorker server (nsrshutdown -a).

       2.     Change to the /nsr directory (cd /nsr).

       3.     Save the temporary resource directory created by nsrize (mv
              res res.save).

       4.     Move the recovered resource directory into place (mv res.R
              res).

       5.     Restart the NetWorker server (cd / ; nsrd).

       6.     After verifying that the recovered resources are valid, remove
              the temporary resource directory (rm -r /nsr/res.save).

       NOTE: The recoverindex command is only used to recover the NetWorker
             server's index.  Use recover to recover a client's index.

       Once recoverindex is run and the media database has been recovered,
       one can use the recover by save set feature to restore entire
       filesystems.  This method can be faster as one does not have to use a
       browser to mark all the files and directories in a file system, an
       operation which can take a long time.  You should only use this
       method for recovery of a filesystem when you can locate a recent save
       set for that filesystem which was saved with level=full.  When
       recovering multiple save sets which are interleaved on media, recover
       will recover all save sets concurrently instead of making a separate
       pass over the media for each save set.  See the recover(8) man page
       for details on running recover by save set.

       If you want to recover whole filesystems, an alternative to using
       recoverindex and recover is to use the scanner command to recover all
       of the files in a particular save set.  The scanner command can also
       be used to print a table of contents for a volume to help you locate
       the correct save set for a filesystem.  Similar to recover by save
       set, you should only use scanner for recovery of a filesystem when
       you can locate a recent save set for that filesystem which was saved
       with level=full.  If using this method of recovery, use of the -x
       option is preferred to simply piping the output through uasm as
       multiple save sets interleaved on the media can be read concurrently
       instead of making a separate pass over the media for each save set.
       For example, the first command is preferred to the second:




Licensed material--property of copyright holder(s)                         2





NSR_CRASH(8)               Legato NetWorker 4.1.1               NSR_CRASH(8)


              scanner -s 16234 -s 16257 -x'uasm -r'
              scanner -s 16234 -s 16257 | uasm -r

       See the scanner(8) manual page for more details.

       If the server is damaged so badly that it will not run at all, you
       will need to follow the manufacturer's instructions for re-installing
       and rebooting a multiuser system.  Once you have the system up and
       running in multiuser mode, you can re-install NetWorker (i.e. extract
       NetWorker from the distribution tape and install it, using
       nsrize(8)) and use recoverindex to rebuild the on-line index.
       Finally, you will want to recover files which previously existed on
       the machine, but which do not exist on the manufacturer's
       distribution tape.  This may include: system files which had been
       customized, a specially tailored kernel, new special device entries,
       locally developed software, and user's personal files.


SEE ALSO
       nsrize(8), nsrlayout(5), nsr(8), recover(8), savegroup(8),
       saveindex(8), scanner(8).




































Licensed material--property of copyright holder(s)                         3


Typewritten Software • bear@typewritten.org • Edmonds, WA 98026