Calculating the written RAW data size
Posted by Sven Koester, Last modified by Sven Koester on 19 December 2018 13:25
We were often asked about tape capacity, and why the capacity printed on the tape differs from the size shown in the PresSTORE Volume table.
The reason is that the size in the volume table is the sum of the saved files' sizes. That does not always correspond to the bytes physically written to the tape.
Here is a method to find out the size of the raw data written to tape, the result gives a more exact view to the physical tape capacity.
The calculation is rather simple: At the end of each job, PresSTORE writes its configuration database to the tape and outlines in the job protocol the ConfigBlkNo, that is the block number, where the configuration can be found.
That block number can be used to calculate the raw data when the block size is known by simply multiplying.
Here is an example. The volume table shows a tape holds 1.74 TB of data:
This is physically an LTO5 tape, from the blue bar in the table one can see it is nearly full.
The last job on that tape shows the config position at block no 14640538:
Now we still need the blocksize, that is in the media pool's advanced options:
Here, the pool size is set to 128 KB, that turned out after some measurement to be the fastest value for our hardware. The default block size for Media pools is 64 KB.
Now we can calculate with the block number and the block size:
14640538 * 128 KB = 1873988864 KB
Divide that by 1024 * 1024 * 1024 to get the value in TB, that is rounded 1.75 TB.
So 1.75 TB (plus a bit for the size of the config database itself) is the raw written capacity until now for this tape. The tape is appendable in the volume table, so it will be able to hold even more.
Please note that is is by chance that the two values 1.74 TB sum of the data sizes and 1.75 TB raw size are that close together, there may be cases where the values differ far more.