GraphEngine icon indicating copy to clipboard operation
GraphEngine copied to clipboard

Questions: use new core to run DistributedHashtable

Open bidiudiu opened this issue 7 years ago • 1 comments

Hi, Sample DistributedHashtable works fine with old version1.0.8482. But when I build the code and get Core 2.0.9328 and use the new core to run sample DistributedHashtable, I encounter some problems: (1) when my two servers are in different AvailabilityGroups, server.start() is stuck at: image Below is my trinity.xml. Is there anything wrong with it? Do I need other settings? image

(2)when my two servers are in the same AvailabilityGroup, "set" and "get" operations aren't working properly. Here's a input-output example: 1 Set("0", "a") 2 Get("0")=> a 3 Set("0", "b") 4 Get("0")=>a 5 Get("0")=> a 6 Get("0")=> b 7 Get("0")=> b 8 Get("0")=> b 9 Get("0")=> b 10 Get("0")=> a 11 Get("0")=> b ... ... (3)In core 2.0.9328, Global.CloudStorage.GetServerIdByCellId(...) is removed and Global.CloudStorage.GetPartitionIdByCellId(...) is available. What's the difference between them?

Thanks a lot!

bidiudiu avatar Mar 28 '18 12:03 bidiudiu

hi @bidiudiu I'll take a look at the DHT example. If the two servers are in the same AG, the default fixed memcloud is not going to hold consistency. You're going to use the GraphEngine.DynamicCluster package to get HA support.

yatli avatar Apr 08 '18 03:04 yatli