QBFS Versus Blockpool Wrapper

Overview 

 

 

 

With DXi 2.0 Software, QBFS is gone. But the Blockpool Wrapper (BPW) has been added.

 

StorNext (SNFS) and Blockpool (BFST) are  the same as they were with Galaxy 1.x. But now, the blockpool wrapper is doing all the work.

 

 

 

The only arrows going into the blockpool wrapper are the protocols and the API. Nobody else talks to BPW. Everyone who wants to talk to the system at the bottom of the architecture should do so through the BPWAPI. That’s where all these other arrows – the rest of DXi – talk to the system, through the API.

 

Notice in the DXi 2.0 Software diagram that no processes are reaching down into StorNext, except for the blockpool wrapper. The blockpool wrapper is the interface between protocols and the data stores (SNFS and BFST). It is essentially an RPC server from the other side of the shared memory (ddup_sm).

 

The blockpool wrapper shields the blockpool for the most part. Replication, for example, has to get information from the blockpool wrapper. If bpgc needs to scan a filesystem, it has to get the tags from blockpool wrapper. The space manager will actually talk to SNFS, but only for low space events.

 

All presentations (such as NFS) are modified to use pcache shared memory. BPW feeds off of this shared memory to act upon the data. It’s inline - data does not touch the disk until the final resting place. BLOB tags are handled by SNFS, not by QBFS.

 

For CIFS/NFS, data written to non-deduplicated shares will be directed by BPW to the underlying filesystem. This allows CIFS/NFS to use the same data path for both types of shares.


What's Next?

Blockpool Data, Files, and Commands >

 

 
 

 



This page was generated by the BrainKeeper Enterprise Wiki, © 2018