AnsweredAssumed Answered

High CPU on target after transfer

Question asked by dsaum on Jul 27, 2011
We've got an alfresco instance running share with WCMQS locally. For a couple weeks we've been doing replication via transfer jobs to another instance internally. In preparation for a launch, we've configured a new instance of Alfresco on an external server and we're attempting to transfer from our internal instance to it.

The transfer seems to work fine (from the source machine's perspective) but on the target, immediately after the transfer ends, the CPU skyrockets (100% use) and stays that way indefinitely (over two hours before we killed it). After a restart, the load is still at 100%.

Note that as part of the same transfer job we are targeting the existing internal instance that has worked in the past and it receives the transfer without a problem.

We enabled debug logging on the target machine and see a large number of messages (tens of thousands) that look something like this (note the Fetched Resource: null).

15:10:32,726 User:System DEBUG [util.transaction.SpringAwareUserTransaction] Began user transaction: UserTransaction[object=org.alfresco.util.transaction.SpringAwareUserTransaction@3479501a, status=0]
15:10:32,730 User:System DEBUG [util.transaction.SpringAwareUserTransaction] Committed user transaction: UserTransaction[object=org.alfresco.util.transaction.SpringAwareUserTransaction@3479501a, status=3]
15:10:32,732 User:System DEBUG [util.transaction.SpringAwareUserTransaction] Began user transaction: UserTransaction[object=org.alfresco.util.transaction.SpringAwareUserTransaction@69c7f04d, status=0]
15:10:32,739 User:System DEBUG [repo.transaction.AlfrescoTransactionSupport] Bound txn synch: TransactionSychronizationImpl[ txnId=d0eff983-899c-4a88-a50b-efc66696c4d3, daos=0, integrity=0, indexers=0, resources={}]
15:10:32,740 User:System DEBUG [repo.transaction.AlfrescoTransactionSupport] Bound resource:
   key: RetryingTransactionHelper.ActiveTxn
   resource: UserTransaction[object=org.alfresco.util.transaction.SpringAwareUserTransaction@69c7f04d, status=0]
15:10:32,740 User:System DEBUG [repo.node.MLPropertyInterceptor] Intercepting method exists using content filter en_US
15:10:32,740 User:System DEBUG [repo.node.MLPropertyInterceptor] Intercepting method exists using content filter en_US
15:10:32,741 User:System DEBUG [repo.service.StoreRedirectorProxyFactory] Redirecting method public abstract boolean org.alfresco.service.cmr.repository.NodeService.exists(org.alfresco.service.cmr.repository.StoreRef) based on store type user://alfrescoUserStore
15:10:32,741 User:System DEBUG [repo.transaction.AlfrescoTransactionSupport] Fetched resource:
   resource: null

The main differences between the two target machines are:

1. Problem machine is 64bit
2. Problem machine does not have share.war deployed
3. Problem machine is not on our internal network and cannot communicate with ActiveDirectory
4. Problem machine is running on an Amazon EC2 instance using Elastic Block Storage as its disk

Any ideas?