Hi,

If the database changes significantly between full backups we could
try lowering the block size on which the hash is computed

For e.g.
/quadstorvtl/bin/vtconfig -m -v VTL_QUA -b 32768

-b (block size) can be one of 0, 4096, 8192, 16384, 32768, 65536,
131072 or 262144
Setting the block size will revert to the default of VTL to determine
the block size. In this case it will be 262144 (256k)

The advantage of reducing the block size is that the likely hood of
finding a duplicate block is increased. Setting this value to the
database block size might be a good idea.
The downside is that the metadata that is stored and the number of
entries in the deduplication tables increases signficantly. For
example for a 256K dedupe block size to hash a 256k incoming block,
the deduplication table entry is around 42 bytes while with a 8k block
size this increases to 1344 bytes. Similar requirements for the
metadata maintained by the virtual tape. So this only makes sense to
lower the block size if the deduplication ratio is high.
The system memory requirements also go up but with the 128 GB memory
size on this system its not a problem.

Please try with two full database backups after setting a lower block
size and then send the diagnostics again.  Also send us the output of
rman 'list backup' for the newly performed backups.

The replication ip addresses are not correctly set.
VTL Replication host should be set to the IP address of the remote VTL
to which data needs to be sent.
On the source VTL ndrecv.conf RecvAddr needs to be removed and on the
destination VTL ndrecv.conf RecvAddr should be set to the interface IP
over which the destination VTL will receive replication data.

On Sat, Jun 8, 2019 at 2:19 AM Jorge Bejarano <[email protected]> wrote:
>
>
> Hello, we only send Oracle database without archives to this vtl.
>
> Regards
>
> El miércoles, 5 de junio de 2019, 3:08:20 (UTC-3), Jorge Bejarano escribió:
>>
>> Hello,
>>
>> I'm taking backups of Oracle Databases with rman and Networker, and the 
>> deduplication ratio does not exceed 4x.
>>
>> This is my rman configuration:
>>
>> CONFIGURE DEFAULT DEVICE TYPE TO sbt;
>> run
>> {
>> allocate channel t2 type 'SBT_TAPE'  format 'b_%u_%p_%c';
>> SEND DEVICE TYPE 'SBT_TAPE' 
>> 'NSR_ENV=(NSR_SERVER=hostname,NSR_CLIENT=hostname,NSR_DATA_VOLUME_POOL="Default",NSR_NO_MULTIPLEX=TRUE,NSR_SAVESET_RETENTION=5
>>  days)';
>> BACKUP filesperset 1 DATABASE;
>> }
>> EOF
>>
>> The backup of the file systems of the same clients with Networker has an 
>> average of 30x.
>>
>>
>> Any advice will be of great help.
>>
>> Thank you!
>
> --
> You received this message because you are subscribed to the Google Groups 
> "QUADStor VTL" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/quadstor-vtl/033409e8-9d42-4182-8ab5-a4467cdf1d1e%40googlegroups.com.

-- 
You received this message because you are subscribed to the Google Groups 
"QUADStor VTL" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/quadstor-vtl/CAPEgkt5b_C8oEYc3Cotb4w0AhWeGDYXe5FJVQEZNob37%2ByAXnA%40mail.gmail.com.

Reply via email to