Lester Caine wrote:

>Jim Starkey wrote:
>
>
>
>>The primary benefit of repositories is that large stuff is stored in
>>generally stable files, drastically reducing the size of the primary
>>database. This has a huge effect on backup, restoration, and failover
>>times.
>>
>>
>
>That sound the right sort of approach. Especially for things like
>document scans, which are always read only, and location need not
>change. Once they have been backed up - no need to do it again - only
>extend the backup when extra pages are added?
>
>
>

Exactly. The customer in question scans everything that comes into the
mailroom into a PDF that goes into the "content store" (Netfrastructure
is a "content store", not a "database"). Repository volumes are
generally read only except for the active volume and are usually open
for read shared until a write is required. For reasons having to do
with replication, repository blobs can be assigned by id without copying
the actual data. This dictates that under normal circumstances
repository blobs are not physically deleted, though there is a special
"expunge" operation to make the sucker actually go away. But in most
cases, the blobs are archival and will clutter up disk platters in
perpetuity. Really old ones will eventually migrate to a juke box, but
I plan to retire first.

--

Jim Starkey
Netfrastructure, Inc.
978 526-1376



[Non-text portions of this message have been removed]