Unsolved
This post is more than 5 years old
3 Posts
0
470
November 20th, 2009 00:00
Content Address Calculation before storing the file in Centera
If you have a client-server architecture, storing a large file (stored in the client side) in Centera through the server, and you want to store the file only if it isn´t stored yet in Centera, is it necessary to send always the file to the server side to calculate its content address, or could its content address be calculated in the client side?
No Events found!
EMCDennis
124 Posts
0
November 24th, 2009 10:00
This would come down to the application and the SDK options enabled. There is FP_OPTION_CLIENT_CALCID
which when enabled causes the client to calculate the ID and checks for existence before sending the data.
If the object already exists and is re-written, Centera has organic processes that are always running that will
eventually find and clean up the redundant copies.
luispucela
3 Posts
0
November 25th, 2009 01:00
What I would like to do is not to send the content, just send the ID calculated on the client. But I need the algorithm used in Centera. If I used FP_OPTION_CLIENT_CALCID, the content would be sent to the Centera, and then calculated its id.
holgerjakob_c0722c
2 Intern
•
337 Posts
0
December 7th, 2009 08:00
Hi
Like Dennis I think that Client side calc id is the only mechanism provided by the SDK to achieve what you want.
As you mention this sends the object to Centera.
How likely are identical files? As you mention they are big ones where I typically see little single instancing happening. If you want to prevent a file that was downloaded to the client from being uploaded again you might want to consider storing downloaded files in another directory than newly created ones. You could also track a hash value of the file on your own and compare that one to hashes already in your database before sending the file to Centera.
Best regards, Holger