[Horizon] Session timeout setting affects large snapshot creation.
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Won't Fix
|
Medium
|
MOS Keystone | ||
7.0.x |
Won't Fix
|
Medium
|
MOS Maintenance | ||
8.0.x |
Won't Fix
|
Medium
|
MOS Maintenance | ||
9.x |
Won't Fix
|
Medium
|
MOS Keystone |
Bug Description
Detailed bug description:
Environment has session_timeout in horizon local_Settings.py set to 900 seconds for security compliance reason. While creating large instance snapshots, in this case 80GB snapshot, the snapshot creation process gets stuck in saving and never completes. Performing the same task via CLI works just fine, but takes ~40min. 40 min is significantly longer than the allotted 900 seconds (15 min) and thus the snapshot fails. We have since increased the session_timeout to 3600 for testing reasons and the snapshot completes just fine after making this change. However, problem is that customer requires 900 second session_timeout to meet security compliance. Requesting to find a way for session_timeouts not to affect ongoing processes via Horizon.
Steps to reproduce:
Create snapshot that takes longer to complete than the session_timeout value.
Expected results:
In theory, this should complete itself even if the session times out.
Actual result:
Snapshot gets stuck in saving state and never finishes.
Reproducibility:
100% if criteria in "steps to reproduce" are met.
Workaround:
Use CLI.
Change session_timeout value to a higher value.
Impact:
Not very high since CLI works fine for customer, but they cannot enforce the session_timeout increase for security reasons.
Description of the environment:
- Operation system: Ubuntu 14.04
- Versions of components: MOS 7.0
- Network model: VxLAN
Token revocation due to Horizon session timeout is an expected behavior, you need to use trusts for a long-running operations, please consult with Keystone team.