Skip to content

ADBDEV-1140 ADB update version to 5.28.3 and update gporca submodule#130

Merged
deart2k merged 1 commit intoadb-5.xfrom
5.28.3-ADBDEV-1140
Nov 9, 2020
Merged

ADBDEV-1140 ADB update version to 5.28.3 and update gporca submodule#130
deart2k merged 1 commit intoadb-5.xfrom
5.28.3-ADBDEV-1140

Conversation

@deart2k
Copy link
Member

@deart2k deart2k commented Nov 9, 2020

No description provided.

@deart2k deart2k requested a review from Stolb27 November 9, 2020 08:40
@deart2k deart2k merged commit 2387d83 into adb-5.x Nov 9, 2020
@deart2k deart2k deleted the 5.28.3-ADBDEV-1140 branch December 21, 2021 09:04
RekGRpth added a commit that referenced this pull request Dec 3, 2025
When the DROP IF EXISTS command is called, it is unconditionally dispatched to
the segments, even if the object doesn't exist on the coordinator. This can
lead to inconsistencies, for example:

- the first session starts a table drop on the coordinator. Even though the
table doesn't exist, the drop is still dispatched to the segments.

- at the same time, the second session creates the table (on the coordinator
and on the segments).

- the first session (on the segments) already sees the table and therefore
deletes it.

Don't dispatch DROP if object doesn't exist on coordinator. Exclude the
drop_rename test from the parallel group because it contains fault-injectors.
Add a new GUC, gp_dispatch_drop_always, for unconditional DROP dispatch even if
the object isn't present on the coordinator, as this functionality is used by
the gpcheckcat utility, for example, in the orphan_temp_table test.

Ticket: ADBDEV-8867
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants

Comments