CVE-2015-3241

What it actually does

The flaw prevents the migration process from terminating when a resized instance is deleted, allowing continued consumption of disk, network, and other resources by the orphaned migration task.

Why it matters to the business

Attackers can exhaust shared cloud resources, leading to degraded performance or outages for other tenants and increased operational costs until the migration tasks are manually cleared.

How an attack unfolds

An authenticated user initiates a resize operation on an instance and immediately deletes the instance while the migration is still in progress; because the process does not stop, the system continues allocating and using resources until manual intervention or restart occurs.

How to fix it

Upgrade OpenStack Nova to a release newer than 2015.1.1 that contains the fix for this migration termination logic; apply the latest stable branch patches and restart affected compute services.

How to confirm the fix worked

Verify the running Nova version is newer than the affected releases, review migration-related code paths for proper deletion handling, and test that deleting an instance during resize cleanly terminates the migration task without residual resource use.

Typical attack path

1. Obtain authenticated access to the OpenStack Compute API
2. Issue a resize request against a target instance
3. Issue a delete request for the same instance during the resize migration window
4. Monitor continued resource consumption by the orphaned migration process

How to detect it

Search Nova compute and scheduler logs for sequences of resize followed immediately by delete operations that leave migration tasks running; look for orphaned libvirt or qemu processes and persistent network or disk allocations after instance deletion.

• nova-compute.log entries showing migration continuing after instance deletion
• libvirt processes persisting with instance UUID after delete API call

What your security team should do

• Prioritize upgrade of all Nova deployments to a fixed release
• Implement API-level rate limiting and monitoring on resize and delete operations
• Add alerting for instances that remain in migrating state after deletion