Release Notes / Version 10.2101
Table Of ContentsThe Content Management Server maintains stable and universally unique identifiers (UUIDs) as specified in RFC 4122 for content items and versions. Existing and newly created contents and versions get a random UUID in addition to existing numeric content IDs and version numbers. UUIDs are not meant to replace existing IDs, and IDs can be used for all purposes as before. Note, that UUIDs are currently only available on the Content Management Server, and not on live servers.
Upgrade:
The
Content Server
automatically adds new database table columns for UUIDs when started. This affects both management and live servers. If the database is Microsoft SQL Server or DB2, then you have to set the configuration property
sql.schema.alterTable=true
once when starting a
Content Server
.
Note, that it is not possible to start a Content Server of a previous release on the database after the new Content Server has adapted the database tables. You should always make a backup of the Content Server database before upgrading. This is even more important for this change.
The
Content Feeder
now feeds the content's UUID and the UUID of its latest version. To this end, the fields
uuid
and
latestversionuuid
were added to the Solr schema in
apps/solr/modules/search/solr-config/src/main/app/configsets/content/conf/schema.xml
. You can invoke the
Content Feeder
's JMX operation
reindexAll
of MBean
AdminBackgroundFeed
with the aspect parameter set to
"uuid"
to set UUIDs of already indexed content. The Search Manual explains re-indexing in general in section "Search Engine | Reindexing | Partial reindexing of Content Feeder indices". It is recommended but not required to perform this re-indexing step.
UUIDs for existing content and versions are generated lazily on the Content Management Server when the content or version is accessed the first time after the upgrade. This entails a small performance penality, but only for the very first access. If you follow the recommendation to let the Content Feeder index UUIDs for existing content as described above, then lazy generation of most UUIDs will happen early as side-effect.
Unified API:
In the Unified API , the following methods have been added for UUID support:
ContentObject#getUuid()
ContentRepository#getContent(UUID)
Content#getVersion(UUID)
ContentCreatedEvent#getUuid()
ContentDestroyedEvent#getUuid()
ContentRevertedEvent#getUuid()
VersionDestroyedEvent#getUuid()
The event classes
ContentCreatedEvent
,
ContentDestroyedEvent
,
ContentRevertedEvent
, and
VersionDestroyedEvent
also got new constructors that take an additional UUID parameter. The existing constructors without UUID parameter have been marked as deprecated.
Furthermore, the existing method
ContentRepository#getContent(String)
can now also be called with a content UUID's string representation to lookup a content by its UUID.
If the
Unified API
is connected to a live server or to a server of a previous release, then methods that would return a
UUID
will return
null
instead.
The Unified API Developer Manual mentions UUIDs in the new section "The Content Repository | UUIDs".
Tools:
cm dump
now also outputs the UUID of contents and versionscm
tools that accept UUIDs as input parameter:dump
,publish
,query
,approve
,destroy
,cleanversions
,serverexport
(CMS-16509)