Test/networking and privacy #4
Conversation
- improve error handling for disperse_data
|
This PR would require to re-test once reliability issues from #3 are resolved. |
| class TestNetworkingPrivacy(StepsDataAvailability): | ||
| main_nodes = [] | ||
|
|
||
| @pytest.mark.parametrize("setup_2_node_cluster", [2], indirect=True) |
There was a problem hiding this comment.
should the param be 2? I saw that the default value is 2
subnet_size = 2
There was a problem hiding this comment.
The original idea is to have 1024 as subnet size, assuming the two cluster members will get assigned repeatedly to multiple subnets and receive 1 replica per each subnet. I wanted to get high volume with 2 node only. I keep this open for now, because the v2 version of subnetworks-assignations which should behave the expected way is not used yet.
And the dispersal endpoint is not reliable to even send two messages repeatedly, and when I tried larger data sample I've got 100% failure rate.
There was a problem hiding this comment.
What I would like to have at the end is efficiency guarding test. Something which will allow some initial efficiency to pass, and check if future versions won't get worse.
PR Details
Set of tests to cover networking and privacy.
Issues reported: