AnsweredAssumed Answered

Strange Migration behaviour

Question asked by corbezzoli on Jun 16, 2010
Hi all,

we are experiencing a weird behaviour with our Alfresco 3.2 Labs installation.
We have planned a migration of our company document files (about 600gb) from a shared drive workstation to Alfresco, using ftp (via wput utility).

Alfresco is using an LDAP authentication, and it's installed on a XEN VM, physically hostend on a SAN. The DBMS is MySql 5.

The copy was not carried out, ending at 16gb of data; for those 16gb of files that are seen on the repo (this is the quota size of the alfresco user used for this copy), we ended up with a full filesystem (about 800gb).
The content store seems to hold more than one (identical) copy per file, and we have no versioning aspect.
 
Now we are facing a double problem: on one hand we need to restore Alfresco, since we have some data on Share wiki and thus we cannot simply wipe out the whole content store, and the deletion of files via explorer, or ftp, is unbelievable slow. Using the web client, we get a timeout and the mysql log reports a transaction error:

CannotAcquireLockException: Hibernate operation: could not execute update query; SQL [delete from alf_usage_delta where node_id=?]; Lock wait timeout exceeded; try restarting transaction; nested exception is java.sql.SQLException: Lock wait timeout exceeded; try restarting transaction

On the other hand, we would like to know why this is happened, in order to go further with this migration.

Could someone shed some light, please?

Thanks in advance

Outcomes