ADSM-L

Re: Useful 'move data' options

1998-07-06 09:21:00
Subject: Re: Useful 'move data' options
From: "Moser, Christian" <Christian.Moser AT SCHERING DOT DE>
Date: Mon, 6 Jul 1998 15:21:00 +0200
     I would appreciate the node option very much.
     With it you can move data of a destroyed node from a non collocated

     tape back to a diskpool in the hierarchy. That can be done while
the
     crashed nodes hardware is recreated and can than improve restore
     throughput highly.

     Kind regards,
     Christian Moser within the Schering Company


____________________________ Antwort-Abtrennung
________________________________
Betreff: Useful 'move data' options
Autor:  ,Bruce Elrick [SMTP:bruce AT WINTERLAND.MAINLAND.AB DOT CA] bei BE2183
Datum:    01.07.98 22:12


Would anyone find the following options to the 'move data' command
useful?

type
   all - all data on volume, default
   backup - backup data
   archive - archive data
   spacemanaged - space managed data
   - allow pairs of the above three types

node
   - specify a list of nodes whose data is to be moved

filespace
   - specify a list of filespaces whose data is to be moved - may
require
   the node option specified

Since migration from sequential storage pools to sequential storage
pools
already moves collocation clusters, the implementation of the latter two

options should be trivial.  The implementation of the first option might

be
more difficult but would be immensely valuable.  I've seen a number of
situations where backup, archive, and space managed data are mixed in
a storage pool (bad idea from a data integrity point of view) and there
is
currently no way to separate the data.

Anyone agree on this?

Cheers...
Bruce
<Prev in Thread] Current Thread [Next in Thread>