/
2.3.5.rst
62 lines (42 loc) · 1.99 KB
/
2.3.5.rst
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
.. _version_2.3.5:
=============
Version 2.3.5
=============
Released on 2018/03/20.
.. NOTE::
If you are upgrading a cluster, you must be running CrateDB
:ref:`version_1.1.3` or higher before you upgrade to 2.3.5.
If you want to perform a `rolling upgrade`_, your current CrateDB version
number must be at least :ref:`version_2.3.0`. Any upgrade from a version
prior to this will require a `full restart upgrade`_.
.. WARNING::
Before upgrading, you should `back up your data`_.
.. _rolling upgrade: https://crate.io/docs/crate/howtos/en/latest/admin/rolling-upgrade.html
.. _full restart upgrade: https://crate.io/docs/crate/howtos/en/latest/admin/full-restart-upgrade.html
.. _back up your data: https://crate.io/docs/crate/reference/en/latest/admin/snapshots.html
.. rubric:: Table of contents
.. contents::
:local:
Changelog
=========
Fixes
-----
- Fixed handling of ``NULL`` values correctly in the cluster tab of the Admin
UI.
- Updated the *Size* label in the table detail tab of the Admin UI to *Size
(Sum of primary shards)*.
- Fixed an issue where ordering by a ``TIMESTAMP`` :ref:`scalar <gloss-scalar>`
would cause the query to get stuck.
- Fixed a potential ``NullPointerException`` when running ``SHOW CREATE TABLE``
with plain type indices.
- Fixed an issue that would cause queries ordered by a ``STRING`` :ref:`scalar
<gloss-scalar>` with possible ``NULL`` values to get stuck.
- Fixed an issue that caused nodes to crash with ``OutOfMemoryException`` when
executing queries with ``ORDER BY`` on large data sets. Instead, the query
circuit breaker is tripped, the query is killed and the node’s health is
unaffected.
- Improved the error handling for a race condition to make sure it doesn't
cause queries to get stuck.
- Improved the behaviour in case the cluster drops below the number of minimum
nodes. The root REST endpoint will no longer timeout and ``sys.checks`` can
still be queried, similar to how ``sys.nodes`` can still be queried.