Released on 2017/07/06.
Note
If you are upgrading a cluster, you must be running CrateDB :ref:`version_1.1.3` or higher before you upgrade to 2.0.4.
If you want to perform a rolling upgrade, your current CrateDB version number must be :ref:`version_2.0.0`. If you want to upgrade from a version prior to this, the upgrade will introduce all of the breaking changes listed for :ref:`version_2.0.0`, and will require a full restart upgrade.
Consult the upgrade notes for :ref:`Version 2.0.0 <version_2.0.0_upgrade_notes>` when upgrading.
Warning
If you're using CrateDB's BLOB storage you should consult the Upgrade Notes.
Before upgrading, you should back up your data.
Table of contents
Due to a bug introduced in :ref:`version_2.0.0` that can cause loss of BLOB data, it is necessary to perform a rolling upgrade if you're running a version greater than or equal to 2.0.0 and using BLOB tables.
Additionally, the number of replicas needs to be set to at least 1
for
all blob tables and you need to make sure that data is fully replicated
before continuing.
Only then you may upgrade one node after each other.
- The
recovery_after_time
node check now fails if it would prevent thegateway.expected_nodes
setting from having any effect.
Fixed
SELECT settings['gateway']['recover_after_time'] FROM sys.cluster
so it returns the correct value.Fixed an issue that caused blob data to be deleted upon node shutdown.
Fix thread-safety issue for :ref:`scalar functions <scalar-functions>`
md5
andsha1
.
information.schema.tables
now returns the default routing hash function if none is present in the table metadata.Fixed an issue that caused an exception to be thrown when using
COUNT(*)
with :ref:`subselects <gloss-subquery>` that return one row, e.g.:SELECT count(*) FROM t1 WHERE id > (SELECT max(col1) FROM t2)Throw proper exception when using the
_raw
column inside the where clause instead of silently ignoring it.