[Deploying Sakai] File System Storage, Disaster Recovery and Archiving

[Chris Ribble]

Hi everyone,

I'm cross posting this to the Sakai Production(originally posted to Dev) 
list to see if any of you fine folks could help me out as well.

For the last 7 years we have been storing all of our binary content in our 
database and it has grown to an uncontrollable size. So large that soon 
recovery from a database fault could take days. So were looking for 
options and I was curious to see what others in the Sakai community are 
doing?

Were considering switching to file system storage and I have a few 
concerns. Right now our database can be restored on a point in time basis. 
This way the entire state of the system reverts back to the specific 
date/time we have selected. With a file system, at least with our restore 
software(TSM) we do not have the same capability. We can restore a file at 
a time or restore the entire file system by revisions. At least thats my 
understanding. How is everyone else handling this?

When converting to file system storage, what strategy has everyone else 
used? Do you shut down and convert? Can you convert while the system is 
being used? I was considering running a separate Sakai instance that would 
be doing the conversion while my production instances served my users. 

How long did your conversion take, how much data did you have? From our 
analysis we have about 550GB of binary data in the database.

On the subject of archiving. What is everyone in the community doing? Do 
you have an "archive" instance of your database and file system? Do you 
package individual sites into flat files and host them on a document 
management system? 

Thanks,
Chris


Chris Ribble
LMS Application Developer
Information Technology Services
University of Windsor
cribble at uwindsor.ca
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/production/attachments/20140124/ca53521d/attachment.html