and others, we tried to increase site & index performance by changing lucene.indexer.mergerMergeFactor parameter in repository setting. ATM this parameter is 13, we started with 8, I think.
We've noticed that performance goes in cycles, like Alfresco merges indexes periodically and performance gets better for a time.
My question is this: is there a way to manually give order to merge index, or is there some parameter which will shorten period between two index merging?
<blockquote> The following properties control the generation and execution of SQL statements to modify the Alfresco schema and data (defaults shown): db.schema.update=true db.schema.update.lockRetryCount=24 db.schema.update.lockRetryWaitSeconds=5 db.schema.update=true:
When this is off, Alfresco will not generate or execute any SQL statements. This can be used to completely avoid any metadata queries between upgrades.
db.schema.update.lockRetryCount and db.schema.update.lockRetryWaitSeconds: If schema updates are on and some modifying SQL has been generated, then the system attempts to get a lock against the database by creating a lock table alf_bootstrap_lock. If the table already exists, the system will wait and repeat the process of metadata examination and script generation. If the lock table cannot be created after all the retry attempts have been exhausted, the system will post a message and exit: </blockquote>
So, if there is in Alfresco DB lock table, Alfresco will try to perform update defined number of time in a loop, then exit.
For some reason, I had alf_bootstrap_lock table in my Alfresco development DB which I used on my VM. When droppped, 4.2.c –> 4.2.f –> 5.0.d upgrades went just fine. Upgrade 5.0.d –> 5.1.g seems OK, but JBoss stops or freezes when Alfresco WAR is started. Probably is something wrong in web.xml or libraries.
