This is an umbrella issue to investigate why the Confluence upgrade to 5.7 takes a long amount of time with certain database configurations or in some rare cases even hangs.

Here is an example log message:

2015-02-05 01:35:26,704 INFO [localhost-startStop-1] [upgrade.upgradetask.attachmentceo.AttachmentIdClashResolver] resolveAttachmentIdClashes No ID collisions were found between the ATTACHMENTS and CONTENT tables.
2015-02-05 01:35:26,717 INFO [localhost-startStop-1] [upgrade.upgradetask.attachmentceo.AttachmentToContentMigrator] doMigration Beginning migration of 6644 attachment entries..

This appears to be affecting SQL Server and Oracle databases that already have latency issues.

Workaround for SQL Server

• SQL Server is known to have this issue when a connection validation query is enabled or when C3P0 (direct JDBC) connection pooling is used.
• There is a bug (CONF-36484) where the preferredTestQuery passed to C3P0 is ignored causing it to issue a "list tables" query which locks up if run while the upgrade task is running.
• The solution is to switch to using a datasource connection with "select 1" as the connection test query.