Insufficient RPC default timeout for pre_live_migration
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Expired
|
Undecided
|
Unassigned |
Bug Description
Pre_live_migration on destination node take a long time if the base image of the instance to be moved is not yet cached and large. The default RPC timeout of 60s is insufficient for this. This would result in a timeout error and migration will be aborted. A typical usecase is an instance spawned off a snapshot for the first time.
However, changing the default timeout globally is not recommended since it's difficult to model the system behaviour changes that can be caused by this. In addition, we have seen that we need timeouts of over 1200s in certain scenarios and this is obviously unsuitable for a global timeout.
It would be good to change the timeout just for the pre_live_migration RPC api and keep it configurable.
tags: | added: compute |
Changed in nova: | |
importance: | Undecided → Medium |
status: | New → Confirmed |
Changed in nova: | |
assignee: | nobody → Yaguang Tang (heut2008) |
Changed in nova: | |
status: | In Progress → New |
assignee: | Yaguang Tang (heut2008) → nobody |
Changed in nova: | |
status: | New → Confirmed |
Changed in nova: | |
assignee: | nobody → jichencom (jichenjc) |
tags: | added: live-migrate |
tags: |
added: live-migration removed: live-migrate |
Changed in nova: | |
assignee: | nobody → lvmxh (shaohef) |
Changed in nova: | |
assignee: | nobody → stgleb (gstepanov) |
Fix proposed to branch: master /review. openstack. org/61728
Review: https:/