Performance Index in Centera

Could you elaborate on what you mean by "performance index"

If I remember correctly the 'performance index' was/is the name used internally in Centera development for the BlobIndex. The BlobIndex is a database that lives on every disk in a Centera cluster and stores information about which fragments are stored locally on that disk, together with additional information that enables faster retrieval of these fragments (hence the name 'performance index'). In addition it also stores information about protection chains (I have fragment X here locally, where is the mirror fragment stored?), and some indexes (query, replication).

Most of the time it plays the role of a cache: when it has the requested information certain operation can be done more efficiently (reading for example). However, for some operations it's vital. Query for example uses an index on top of the BlobIndex.

I think for a 'normal' user it's something not to worry about: it's an internal datastructure. It's only when it fails (database corruption, disk failure) that it becomes noticeable because some operations will either be slower (reads for example can become slower) or not working (query, in case 2 or more databases at once are unavailable).

To my knowledge we have never published any details on the internal workings of the Centera outside of engineering.  Also there is nothing that a developer can use via the API to influence the internal indexing.

Hi Paul,

Thank you for the reply. However, let me know if there is any document available as such, which I can read, for my knowledge.

Regards,

Supratik Khandelwal

Sr. Technical Instructor

EMC Corporation