Skip to content
This repository has been archived by the owner on Mar 26, 2020. It is now read-only.

Unified Volume Create API and mount bricks on Glusterd2 start #851

Closed
aravindavk opened this issue Jun 4, 2018 · 6 comments
Closed

Unified Volume Create API and mount bricks on Glusterd2 start #851

aravindavk opened this issue Jun 4, 2018 · 6 comments
Assignees
@aravindavk
Labels
FW: Device Management FW: Volume Management GCS/1.0 Issue is blocker for Gluster for Container Storage priority: high

Comments

@aravindavk
Copy link
Member

aravindavk commented Jun 4, 2018 •
edited
Loading

Based on the comments in PR #661, listed the fixes required for Smart Volume Create API

  • Single REST API /volumes and auto provision bricks if size is specified.
  • Volume name as optional parameter
  • Mount bricks on Glusterd2 start or Volume Start
  • Save device/lv details in brick.MountInfo(This structure is added by Snapshot feature, useful here also. We can even skip custom fstab logic since this is sufficient for mounting bricks on Glusterd2 start)
  • Add omitempty for optional API parameters
@aravindavk aravindavk self-assigned this Jun 4, 2018
@aravindavk aravindavk mentioned this issue Jun 4, 2018
Merged
@aravindavk aravindavk mentioned this issue Jun 11, 2018
Merged
@atinmu atinmu added priority: high Gluster 4.2 GCS/1.0 Issue is blocker for Gluster for Container Storage labels Jun 22, 2018
@aravindavk aravindavk added this to the GCS-Sprint1 milestone Jun 25, 2018
aravindavk added a commit that referenced this issue Jun 25, 2018
@aravindavk
Auto generate Volume Name if not specified in Request
31f8290 
Updates: #851
Signed-off-by: Aravinda VK <avishwan@redhat.com>
@aravindavk
Copy link
Member Author

I will open new issues for each of the following tasks and move out of GCS-Sprint1

Auto distribute count: If Distribute count is not specified and disk space is not enough to create volume, try again with increased distribute count till all the bricks can be planned(till max distribute count)

Arbiter brick size calculation: Introduce configurable option for average file size with good default value. Apply the formula while planning arbiter brick

Global defaults should be configurable: Default replica count, Default Snapshot reserve factor, etc should be configurable via API/Config

Add support for multiple filesystem types(Default to xfs)

@aravindavk aravindavk mentioned this issue Jul 9, 2018
Merged
This was referenced Jul 12, 2018
Open
Open
Closed
Closed
Open
Open
@atinmu atinmu modified the milestones: GCS-Sprint1, GCS-Sprint2 Jul 12, 2018
@atinmu
Copy link
Contributor

atinmu commented Jul 12, 2018

Moving this to Sprint 2 as there's good amount of work required to complete all of the issues referenced.

@aravindavk
Copy link
Member Author

Moving this to Sprint 2 as there's good amount of work required to complete all of the issues referenced.

Patch #973 posted for the tasks other than the new issues opened. It can be merged if somebody reviews it.

@aravindavk
Copy link
Member Author

@atinmu updated status in first comment.

@aravindavk aravindavk changed the title Intelligent Volume Provisioning improvements Unified Volume Create API and mount bricks on Glusterd2 start Jul 13, 2018
@aravindavk
Copy link
Member Author

Updated title and description of issue

kshlm pushed a commit that referenced this issue Jul 18, 2018
@aravindavk @kshlm
Mounting auto provisioned bricks on glusterd2 start
7603e0d 
On `glusterd2` restart, it checks the list of auto provisioned bricks
and mounts if not mounted already.

Updates: #851
Signed-off-by: Aravinda VK <avishwan@redhat.com>
Madhu-1 pushed a commit to Madhu-1/glusterd2 that referenced this issue Jul 19, 2018
@aravindavk @Madhu-1
Mounting auto provisioned bricks on glusterd2 start
4f9dde7 
On `glusterd2` restart, it checks the list of auto provisioned bricks
and mounts if not mounted already.

Updates: gluster#851
Signed-off-by: Aravinda VK <avishwan@redhat.com>
Madhu-1 pushed a commit to Madhu-1/glusterd2 that referenced this issue Jul 19, 2018
@aravindavk @Madhu-1
Mounting auto provisioned bricks on glusterd2 start
00320d2 
On `glusterd2` restart, it checks the list of auto provisioned bricks
and mounts if not mounted already.

Updates: gluster#851
Signed-off-by: Aravinda VK <avishwan@redhat.com>
@aravindavk
Copy link
Member Author

Closing this issue since #973 is merged.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees

@aravindavk aravindavk

Labels
FW: Device Management FW: Volume Management GCS/1.0 Issue is blocker for Gluster for Container Storage priority: high
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@aravindavk @atinmu