-
Type:
Bug
-
Status: Closed (View Workflow)
-
Priority:
Critical
-
Resolution: Won't Do
-
Affects Version/s: 1.7.0
-
Fix Version/s: None
-
Component/s: None
-
Labels:
-
Story Points:8
-
Epic Link:
-
Sprint:Junco Sprint #2 - Platform, Bug Sprint, K Sprint #1 - Platform, K Sprint #2 - Deployment
ResourceManager requires about 0.5/0.7 seconds to allocate a VLAN, from the moment it is requested. This happens on a 2 core machines with 4 gig of memory.
This causes for example to take more than 5 seconds to install a p2p intent with vlan encapsulation in a 4/5 switches linear topology, with no other intents or specific applications already installed (meaning the system is not already overloaded).
Usually intents get installed in few ms to get installed.
- is duplicated by
-
ONOS-5279 Resource reservation takes too long in multi node cluster
-
- Closed
-
- relates to
-
ONOS-6112 Analyze fundamental issue without update semantic and potential design choices
-
- In Progress
-
-
ONOS-6152 Build state machine backend for ResourceStore
-
- Resolved
-
-
ONOS-4013 A lot of consistent map accesses when unregistering resources
-
- Closed
-
-
ONOS-6227 Pessimistic locking option for transactions
-
- Open
-
-
ONOS-5910 Studying possible solution for conflicts in resource reservation
-
- Closed
-
-
ONOS-6220 Implement distributed lock primitive
-
- Closed
-
-
ONOS-6226 Refactor transactions to improve caching and isolation
-
- Closed
-
1.
|
Identifying the root cause |
|
Resolved | Shimizu Sho |