Enable multi-database creation in the database_objects resource
Overview
We still need to add multiple database creation in the database objects resource as seen at https://gitlab.com/gitlab-org/omnibus-gitlab/-/blob/b80d1237d8a8e38a58bbb230a7ba8e3fd1a17644/files/gitlab-cookbooks/gitlab/resources/database_objects.rb
Deliverables
- If multiple databases are specified, then the
database_objects
resource will create them. - Test installation of application database
- Single Node Database for the application
- Multi-node database for the application
- Geo replication single node database
- Geo replication multi-node database
- Test installation of Geo tracking
- Single Node database
- Multi-Database database
Designs
- Show closed items
Blocks
- gitlab #370798
- gitlab #368728Backlog
Is blocked by
Activity
-
Newest first Oldest first
-
Show all activity Show comments only Show history only
- Robert Marshall added to epic &7394
added to epic &7394
- Robert Marshall mentioned in issue #6434 (closed)
mentioned in issue #6434 (closed)
- Robert Marshall marked this issue as related to #6577 (closed)
marked this issue as related to #6577 (closed)
- Robert Marshall mentioned in epic &7394
mentioned in epic &7394
- Maintainer
@rmarshall Duplicate of #6312 (closed) ?
Collapse replies - Developer
@rmarshall @twk3 Has there been any discussion about a milestone for this issue?
It's currently a blocker for some work we're doing in &6458 (closed). @dbalexandre would be willing to take it on %14.9 if there are no objections.
- Contributor
@nhxnguyen @dbalexandre that would be great. We haven't gone through and schedules these breakdowns from the epic yet. And we wouldn't likely be getting to them until 14.10 at the earliest ourselves.
1 - Developer
- Developer
After digging into the omnibus code, I figured out this is not a blocker for some work we're doing in &6458 (closed). So, I'll unassign myself and focus on #6577 (closed).
1
- Balasankar 'Balu' C mentioned in issue #6193 (closed)
mentioned in issue #6193 (closed)
- Michael Kozono marked this issue as related to gitlab#351946 (closed)
marked this issue as related to gitlab#351946 (closed)
- Douglas Barbosa Alexandre mentioned in issue gitlab-org/geo-team/discussions#5029 (closed)
mentioned in issue gitlab-org/geo-team/discussions#5029 (closed)
- Nick Nguyen changed milestone to %14.9
changed milestone to %14.9
- Nick Nguyen added 1 deleted label
added 1 deleted label
- Douglas Barbosa Alexandre assigned to @dbalexandre
assigned to @dbalexandre
- Michael Kozono added workflowin dev label
added workflowin dev label
- Developer
@dbalexandre This came up in the scheduling call, just wondering if there is any update on this issue?
Collapse replies - Developer
No update yet. This issue is on top of my to-do list for next week.
1
- Fabian Zimmer mentioned in issue gitlab-org/geo-team/discussions#5031 (closed)
mentioned in issue gitlab-org/geo-team/discussions#5031 (closed)
- Douglas Barbosa Alexandre removed the relation with #6577 (closed)
removed the relation with #6577 (closed)
- Douglas Barbosa Alexandre changed milestone to %Backlog
changed milestone to %Backlog
- Douglas Barbosa Alexandre removed 1 deleted label
removed 1 deleted label
- Douglas Barbosa Alexandre unassigned @dbalexandre
unassigned @dbalexandre
- 🤖 GitLab Bot 🤖 added [deprecated] Accepting merge requests label
added [deprecated] Accepting merge requests label
- Douglas Barbosa Alexandre removed the relation with gitlab#351946 (closed)
removed the relation with gitlab#351946 (closed)
- Thong Kuah mentioned in issue gitlab#370611 (closed)
mentioned in issue gitlab#370611 (closed)
- Developer
We will need this issue for &7509 /cc @fzimmer @nhxnguyen
Collapse replies - Developer
Thanks @tkuah. Do you have a sense of when this could become a blocker for &7509?
@twk3 @dorrino Is this something groupdistribution might be able to take on in an upcoming milestone? We're aiming to support decomposition for self-managed customers in 16.0. /cc @gitlabbarry
- Developer
Hey @nhxnguyen due to FedRAMP we will not be able to pick this up before 15.6 at the earliest. Is this okay?
- Developer
Do you have a sense of when this could become a blocker for &7509?
The minimum requirements for &7509 is two connections to same database. This is planned to be the default in %16.0.
When that happens, I imagine some (not all) customers will want ability to use two separate databases as an additional requirement. That is as yet unscheduled so unsure.
- Contributor
We are currently working this epic &7814 (closed) to fix/improve handling of multiple dbs in HA setups which is impacting how we can setup current installs. We will need to finish that first before being able to look at the epic this issue is a part of.
1 - Developer
due to FedRAMP we will not be able to pick this up before 15.6 at the earliest. Is this okay?
From the comments above, it seems that would be ok and we can follow up around then.
- Nick Westbury mentioned in merge request gitlab-environment-toolkit!781 (closed)
mentioned in merge request gitlab-environment-toolkit!781 (closed)
- Thong Kuah marked this issue as related to gitlab#370798 (closed)
marked this issue as related to gitlab#370798 (closed)
- Thong Kuah removed the relation with gitlab#370798 (closed)
removed the relation with gitlab#370798 (closed)
- Thong Kuah marked this issue as related to gitlab#370798 (closed)
marked this issue as related to gitlab#370798 (closed)
- Thong Kuah marked this issue as blocking gitlab#368728 (closed)
marked this issue as blocking gitlab#368728 (closed)
- Balasankar 'Balu' C mentioned in issue gitlab#368728 (closed)
mentioned in issue gitlab#368728 (closed)
- Maintainer
Linking !5492 (merged), which was somthing I opened back when we started work on decomposed databases. But since it wasn't needed at that time, I didn't go much further.
- Developer
Thanks @balasankarc !
Based on gitlab#368728 (comment 1165233903), we need this to be enabled as a dependency of enabling decomposition for self-managed customers. @plu8 @dorrino Is this something we can schedule before %16.0 ? /cc @fzimmer @gitlabbarry
Edited by Thong Kuah Collapse replies - Maintainer
FWIW, I resurrected !5492 (merged), and from a quick testing, it seems to be working fine.
- Balasankar 'Balu' C mentioned in merge request !5492 (merged)
mentioned in merge request !5492 (merged)
- Balasankar 'Balu' C marked this issue as blocked by gitlab-qa#695 (closed)
marked this issue as blocked by gitlab-qa#695 (closed)
- Balasankar 'Balu' C mentioned in issue gitlab-org/distribution/team-tasks#1131 (closed)
mentioned in issue gitlab-org/distribution/team-tasks#1131 (closed)
- DJ Mountney changed milestone to %Next 1-3 releases
changed milestone to %Next 1-3 releases
- DJ Mountney removed For Scheduling label
removed For Scheduling label
- DJ Mountney added Deliverable label
added Deliverable label
- DJ Mountney removed Deliverable label
removed Deliverable label
- DJ Mountney added For Scheduling label
added For Scheduling label
- DJ Mountney mentioned in epic &6159
mentioned in epic &6159
- Developer
When it comes to initializing databases, the way that the problem is framed this is limited to support multi-database Rails application. Registry is a non-Rails stakeholder that can benefit from multi-database support and comes with its own database migration commands.
I am not trying complicate this. Simply highlighting the requirement so that the implementation can accomodate non-Rails applications when needs be.
1 - Hossein Pursultani mentioned in issue gitlab-org/distribution/team-tasks#606 (closed)
mentioned in issue gitlab-org/distribution/team-tasks#606 (closed)
- Jaime Martinez mentioned in issue container-registry#1102 (closed)
mentioned in issue container-registry#1102 (closed)
- Robert Marshall closed with merge request !5492 (merged)
closed with merge request !5492 (merged)
- Robert Marshall mentioned in commit 3abb3ee9
mentioned in commit 3abb3ee9
- 🤖 GitLab Bot 🤖 added workflowcomplete label and removed workflowin dev label
added workflowcomplete label and removed workflowin dev label
- Balasankar 'Balu' C mentioned in commit ce43cc04
mentioned in commit ce43cc04