Datastax Cassandra Administrator Certification Questions and Answer (Pratice Questions and Dumps)

Question-: Hinted handoff is repair mechanism when
A. Read fails on a node
B. Write fails on a node
C. Both read and write fails on a node
D. In case of new node is added to cluster. It will be SyncUP using Hinted Handoff

Answer: B
Exp: While data is being written to a node and that node become unresponsive while data is being written. Because of either H/W, network or overloaded node issue. If the failure detector marks a node as down and hinted handoff is enabled in the Cassandra.yaml file, then all the missed writes are atored by coordinator node for a period of time. Hints are stored in local directory on each node for improved replay.
Hints are flushed to disk every 10 seconds, which reduces the staleness of the hints. When Gossip discovers that a node is online again, then coordinator node replays each remaining hint to write the data on the newly-returned node then hint file would be deleted. If node is down for more than configured time then coordinator node would stop writing new hints.

Admin only



Question-: You have a Cassandra cluster with the two nodes (X and Y) with replication factor is and consistency level is also one. You are writing one row named row key K. Here node X is coordinator node before writing the data node-x goes down, which of the following statement is correct?
A. Node X will not be able to store hint.
B. Node Y has not received data as coordinator node, hence it can not write.
C. Coordinator node will return an UnavailableException error.
D. None of the above

Answer: A,B,C

Explanation: Suppose cluster has two nodes X and Y with consistency level set a 1 as well as replication factor as 1. Now here node X is a coordinator node and goes down before K1 row is written. As node X is a coordinator node is down, it can not store the hint as well. In this case the coordinator will return an “UnavailableException� error. With the write request failed and hints are not written.

Hence, in the cluster we should have enough nodes as well as well defined replication factor. Like 3 nodes with replication factor as 3. So all 3 nodes will get the write request. And if any of the node is down, once of the node will store the hint for that node and when node comes back then hint can be replayed on that node.

Admin Only



Question-: Read repair does not propagate expired tombstones, nor does it consider expired tombstones when actually repairing data. And during read repair if there is tombstone data that has not been propagated to all replica nodes before gc grace expired, this data may continue to be returned as live data?
A. True
B. False

Answer: A
Exp: Read repair does not propagate expired tombstones, nor does it consider expired tombstones when actually repairing data. And during read repair if there is tombstone data that has not been propagated to all replica nodes before gc grace expired, this data may continue to be returned as live data

Admin Only

Related Questions

---

Question-: Which all of the below are correct for the “nodetool tablehistograms� command?
A. It provides the current table level statistics.
B. It provide the detail of an individual table read and write latency since last 15 minues.
C. It provide the information about the entire clutser.
D. It provides the latencies of a node.
E. It provides all the tables name which are not in sync across the cluster


Question-: Please map the followings
A. logback.xml
B. system.log
C. debug.log
D. gc.log

1. You can set the log levels and location of the log files in this.
2. In this you can check when the nodes leaving and joining
3. Access Mostly Uused Products by 50000+ Subscribers
4. You can check the application pause time in this log


Question-: Which of the following information you can check with the “system.log� ?
A. All the gc pauses
B. When the nodes are leaving and joining
C. Metadata ownership change information
D. Index redistribution
E. Which node was attacked by hacker?


Question-: Which of the following are the causes for GC pauses in Cassandra database?
A. lack of JVM space
B. Mistuned JVM GC algorithm
C. Slow disk
D. High volume of data
E. high number of nodes in Cassandra cluster



Question-: While analyzing the Cassandra cluster you found that one of the nodes in the ring is responding quite slow to query. Hence, you decided to debug the system and wanted to enable the log level to Debug only on that particular node. And once analysis is done, you wanted to reset the log level to INFO. Which of the following option you would use, so that you don’t have to restart the cluster?
A. You would be changing the log level in logback.xml file. Which would be effective for next query
B. You would be updating the log level by changing the JVM options.
C. You will be using the nodetool setlogginglevel option.
D. It is not possible, without node restart


Question-: You have setup Splunk process as well as Geneos alert system to analyze the logs. And they both can read the logs from a particular directory only. Now you need to make sure that the logs can be created in that directory only. How can you achieve this?
A. You would be providing the log path at Cassandra.yaml file
B. You will be providing the log path using the jvm.options
C. You would be updating the log4j.properties file
D. You would be updating the logback.xml file