I'm using an AWS DAX cluster of 3 nodes of dax.r4.xlarge node type. When I'm running my spark application from EMR cluster, it is always fetching values from dynamodb table. Even if I run the same application on same set of key, it is querying dynamodb table. In the DAX cluster metrics I see 0 cache hits and misses.
Asked
Active
Viewed 731 times
0
-
Are you querying the dax endpoint, or Dynamo? – hephalump Feb 26 '20 at 19:44
-
Dax endpoint with client without a session – Akshay Feb 27 '20 at 14:01
-
@hephalump Thanks for the help. I fixed it. Added the answer below. – Akshay Feb 28 '20 at 13:32
1 Answers
3
I found the mistake. Initially I was hitting DynamoDB directly and was using consistent reads by defining get item input parameter as:
ConsistentRead: aws.Bool(true)
When I switched to DAX client, I did not remove it. Hence it was always redirecting all the requests to Dynamodb table to keep the reads consistent.
Removing this worked. I can see cache hits and misses in DAX metrics now.
Akshay
- 65
- 7