Guest status doesn't take heartbeat timestamp into account
Bug #1512440 reported by
Mat Lowery
This bug affects 2 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack DBaaS (Trove) |
New
|
Undecided
|
Unassigned |
Bug Description
There are a variety of ways that the guest agent can stop sending heartbeats (or the controller can stop receiving them):
* guest agent crashes
* VM is stopped/
* incompatible heartbeat RPC call (such as happens when the guest and controller are on different Trove versions)
Shouldn't the status derivation code consider "freshness" of the heartbeat?
Changed in trove: | |
assignee: | Amrith Kumar (amrith) → nobody |
To post a comment you must log in.
This was discussed at summit so I'm assigning to myself ... Mat, please let me know if you are already working on a fix and you can reassign to yourself.