When using Courier 2.5 to push content from staging to live we are getting the following error during Extraction, anyone else had something similar and know what the problem may be?
System.Web.Services.Protocols.SoapException: System.Web.Services.Protocols.SoapException: Server was unable to process request. ---> NHibernate.TransactionException: Transaction not connected, or was disconnected at NHibernate.Transaction.AdoTransaction.CheckNotZombied() at NHibernate.Transaction.AdoTransaction.Rollback() at Umbraco.Courier.Core.PersistenceManager.Rollback() --- End of inner exception stack trace --- at System.Web.Services.Protocols.SoapHttpClientProtocol.ReadResponse(SoapClientMessage message, WebResponse response, Stream responseStream, Boolean asyncCall) at System.Web.Services.Protocols.SoapHttpClientProtocol.Invoke(String methodName, Object[] parameters) at Umbraco.Courier.RepositoryProviders.WebServiceProvider.RepositoryWebservice.Rollback(String sessionKey, String username, String password) at Umbraco.Courier.RepositoryProviders.CourierWebserviceRepositoryProvider.Rollback(String sessionKey) at Umbraco.Courier.Core.Extraction.RevisionExtraction.Extract() at Umbraco.Courier.Core.Tasks.ExtractionTask.Run() at Umbraco.Courier.Core.TaskManager.(IRevisionTask value) 15/12/2011 18:54:03
Does the extraction complete even when this error is there?
Basicly it looks like it was trying to roll back a transaction but the transaction had already ended before it could do that.
Transactions can complete in 2 ways, either they complete and are committed to the DB, or they fail and are rolled back, if they time-out, connection is dropped etc, they should also rollback, or get dropped by the destination database.
Very bizarre, the transfer was definitely failing as these were new content nodes that were being migrated across and nothing was being created on live.
We carried out a manual package, copy, and extraction on one of the content nodes that was failing and this succeeded however since then everything is working fine! Each content node with sub pages that was originally failing is now migrating across without problem.
Obviously we will update you if it happens again but for the time being everything is working ok (hate intermittent problems).
Ok, keep me posted, if you see it happening again,
Nic also wrote me via email, my reply below for future reference:
"So from the top of my head, an issue could be the destination servers application pool restarting, as that will kill the transaction and the session nhibernate is using Basicly when you start a deployment from one location to another Courier creates 2 session tokens, one on each site, so it can ping status's back and forth and resumea nhibernate session, instead of having one long running request, it splits things into smaller chunks So if your destination site looses it's app pool or state in some way, Courier looses the transaction and session it was using. This is what I suspect is happening here, as ityou said it appears it works fine when doing a local deploy"
Courier 2.5 extraction failing
When using Courier 2.5 to push content from staging to live we are getting the following error during Extraction, anyone else had something similar and know what the problem may be?
Does the extraction complete even when this error is there?
Basicly it looks like it was trying to roll back a transaction but the transaction had already ended before it could do that.
Transactions can complete in 2 ways, either they complete and are committed to the DB, or they fail and are rolled back, if they time-out, connection is dropped etc, they should also rollback, or get dropped by the destination database.
Very bizarre, the transfer was definitely failing as these were new content nodes that were being migrated across and nothing was being created on live.
We carried out a manual package, copy, and extraction on one of the content nodes that was failing and this succeeded however since then everything is working fine! Each content node with sub pages that was originally failing is now migrating across without problem.
Obviously we will update you if it happens again but for the time being everything is working ok (hate intermittent problems).
Ok, keep me posted, if you see it happening again,
Nic also wrote me via email, my reply below for future reference:
is working on a reply...