topic Disaster recovery CRISIS in Alfresco Archive

Disaster recovery CRISIS

globaladvisors_ — Wed, 27 Mar 2013 11:43:05 GMT

HelloThanks in advance for any help. This is an absolute crisis for us.We have been running Alfresco 4.2b on Debian and Postgres - it has been stable and reliable.Our server crashed last week and we have been attempting to restore from hot-backups via Bacula.We restored the full Alfresco directory a

Re: Disaster recovery CRISIS

michaelböckling — Wed, 27 Mar 2013 12:23:00 GMT

This looks like an inconsistent backup. Are you sure your DB, ContentStore and Index have been backed up and restored correctly? Are you using Lucene or Solr?
You could try to start with a fresh index…

Re: Disaster recovery CRISIS

globaladvisors_ — Wed, 27 Mar 2013 12:34:14 GMT

Hi Michael

Thanks for the response. We're using Solr. I followed the instructions for a hot backup restore to the letter.

I have set index recovery to full. But Alfresco never gets there. I have additionally restored from the solrBackup directory to the solr directory to avoid index corruption.

You will see from the log that it tries to insert into a table:

<blockquote>Error updating database. Cause: org.postgresql.util.PSQLException: ERROR: duplicate key value violates unique constraint "protocol"
Detail: Key (protocol, identifier)=(system, system) already exists
### The error may involve alfresco.node.insert.insert_Store-Inline
### The error occurred while setting parameters</blockquote>

Re: Disaster recovery CRISIS

globaladvisors_ — Wed, 27 Mar 2013 20:09:41 GMT

Hi

I discovered the underlying problem. In setting up the new server, Postgres had upgraded to version 9.1.8-1 on the latest Debian Testing.

Alfresco's JDBC is not comptiable with this version. I downgraded to the version shipped as part of the Alfresco 4.2.c bundle.

I believe other SQL errors being reported might be due to the same issue.

Well. Three days of my life gone, more grey hair and lost sleep. Thought our backups had failed…