Skip to content

S3 fencing to suspend old Primary from writing to S3.#2417

Open
pruthvitd wants to merge 1 commit intoRamenDR:mainfrom
pruthvitd:s3-fencing
Open

S3 fencing to suspend old Primary from writing to S3.#2417
pruthvitd wants to merge 1 commit intoRamenDR:mainfrom
pruthvitd:s3-fencing

Conversation

@pruthvitd
Copy link
Copy Markdown
Member

@pruthvitd pruthvitd commented Feb 16, 2026

Implements failover marker mechanism to coordinate S3 writes between clusters during DRAction failover operations. Prevents data corruption from concurrent writes by suspending old primary's S3 writes when failover marker is detected.

Key changes:

  • Add failover marker creation/removal in DRPC
  • Add marker check in VRG before S3 writes
  • Add marker phase tracking (initiated/recovering/completed)

Tested the failover scenario with both clusters online, and it successfully prevented split‑brain.

Fixes: #250

Implements failover marker mechanism to coordinate S3 writes between
clusters during DRAction failover operations. Prevents data corruption
from concurrent writes by suspending old primary's S3 writes when
failover marker is detected.

Key changes:
- Add failover marker creation/removal in DRPC
- Add marker check in VRG before S3 writes
- Add marker phase tracking (initiated/recovering/completed)

Tested the failover scenario with both clusters online, and it successfully prevented split‑brain.

Code adjustments were made to satisfy linter requirements.
    AI assistance from Microsoft 365 Copilot (with enterprise data protection
    in Microsoft 365 Copilot and Microsoft 365 Copilot Chat) was used for suggestions only.
    All changes were reviewed, adapted, and tested by the author.

    Co-authored-by: M365 Copilot(Enterprise data protection enabled)

Signed-off-by: pruthvitd <prd@redhat.com>
@raaizik
Copy link
Copy Markdown
Member

raaizik commented Feb 24, 2026

LGTM

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.

During failover situation, prevent multiple primary clusters concurrently writing to S3

3 participants