HANA Flashcards
Where can you check if hardware on a customer’s HANA system is supported?
Check the HANA hardware directory
If I upgraded my HANA database from revision 125.10 to 125.35, what kind of release would I call 125.35?
Maintenance Revision
What happens if a customer uses more memory than their license allows?
Depends on type of license (temp/perm, enforced/unenforced)
What does the sapstartsrv service do?
1) Starts the Daemon.
2) Collects the log files and trace files if the database crashes.
If the global allocation limit for memory is nearly at capacity, and we then see a large memory request that would exceed the limit, what would happen
1) Out of memory dumps (OOM) would occur.
2) HANA will unload tables and columns from the main memory to make space.
What are savepoints and what are they used for?
1) Used to recover the database to a consistent state without needing redo logs.
2) A savepoint is overwritten when the next savepoint is successfully completed.
3) They can be used to speed up a restart of HANA DB.
If a transaction causes a commit statement that inserts millions of new records into a column store table and then an Auto-Merge occurs on this table, what does this sequence of events look like?
1) A savepoint occurs and HANA flushes all the changed data from memory to the data volume.
2) During the write operation, log entries are written to the log buffer in memory.
3) During the commit, log entries are flushed to disk synchronously.
If you update a record in a row store table using SQL in the SQL editor in HANA studio and the Auto Commit feature is turned ON, which component does this write operation occur in for MVCC?
Transactional Version Memory (Write operations create new versions in TVM, visible version then moved from TVM to persisted segment based on commit ID)
What is the best way to partition a table over 3 different worker nodes that breaks the table into 3 ranges. For example year 2019 - 2020, 2021 - 2022, 2022 - 2023?
Partition the table with multi-level partitioning using HASH at the end of the 1st level for load balancing and RANGE at the 2nd level for time based partitioning
In a scale out system with 1 master node and 2 worker nodes and 1 standby node, how many index servers are there?
1 master and the 2 workers index servers are active.
If a customer wants to migrate their BW system to an on-premise HANA solution, but have a low budget, what should they do to reduce costs and use their existing software instead?
Tailored Datacenter Integration (TDI)
What types of compression are used in HANA?
Clustered
Prefixed
Indirect
The different services startup in specific sequences. What is the name of the process that decides this sequence?
Runlevel
When checking the table definition for a distributed table in a scale out system, and select the Runtime Information tab in HANA studio, what can we see?
1) Partition specification.
2) Number of partitions.
3) Delta size for column store tables.
How do you check if there are indexes on a table or for the data in the table?
1) In the table definition tab.
2) In the table content tab.
If you check the expensive statements list in HANA studio and find that there is no information, why might this be?
The expensive statements trace is not turned on.
What are the prerequisites to access HANA cockpit?
1) The web browser must support SAPUI5.
2) If your HANA database is on SPS10 or higher you can logon as the SYSTEM user.
3) The DB user needs to be assigned the sap.hana.admin.roles::monitoring or sap.hana.admin.roles::administrator roles
What can you see in the catalog folder in HANA studio?
Tables, Views and Stored Procedures.
A HANA DB is receiving many low priority alerts related to the number of records in column store tables. There are no obvious reasons for these alerts. What would be the next logical thing to do?
You increase the thresholds for the occurrence of this type of alert to more meaningful value
How do you collect runtime environment dumps on HANA DB?
Configure the collection of diagnostic information in the diagnostic files tab of HANA studio.
What is the best way to analyse a poorly performing SQL statement with 1000s of evictions per hour?
Check if bind values are used in the WHERE clause of the statement
If you want to capture all SQL statements that exceed a defined threshold, what kind of trace would you switch on for HANA?
Expensive statement trace.
If the global statement memory limit is 100 GB and the user statement memory limit is 200 GB, and a user wants to execute a report that will consume 150 GB, what will happen?
The report will run successfully as the user statement limit takes priority over global statement memory limit
If the statement memory limit on HANA DB is set to 100 GB, and you sometimes see queries are terminating even though there is low overall memory usage, how would you allow a statement to continue when overall memory usage is below a certain percentage?
Activate parameter statement_memory_limit_threshold
If a HANA DB is running slow when you logon, what is important to check?
1) Check the threads tab to see if any threads are blocked.
2) Check the threads tab to see if there are any long running threads.
3) Check the load graph to see if there are any CPU spikes.
If you see a lot of cache evictions per hour in the SQL cache on all hosts and you also see a used percentage of 99, and each cache is 2GB and the ratios are above 95%, what might be happening?
The cache might be too small due to the large number of statements so they are being evicted to make space.
What KPIs are important for row store reorganizations when a row store table has become fragmented?
1) Row store size should be greater than 10 GB before a reorg.
2) Free space ratio should be greater than 30% before a reorg.
Where can you see a graphical representation of statement executions on HANA studio?
Plan Visualizer. (PlanViz
What are the 2 lock types for application transactions in HANA?
Object, Record.
If there are 2 systems configured for replication, a primary and secondary system, and the secondary system is suddenly unavailable, then we see that primary system is now blocked, what type of replication is in place?
Synchronous Full Sync
If node 3 crashes in a scale out system with a configuration of 1 master and 3 worker nodes and 1 standby node, a failover is triggered and standby node takes over for node 3. What is the sequence of events that lead to the discovery of the crashed host and the failover to the standby node?
Both the name server and the daemon on the crashed host fail to reply to the pings from the master name server.
If a customer implements system replication between a primary and secondary system, and they want to preload the column store tables on the secondary, what would be the pros and cons?
Pro - Shorter ramp up during a failover to secondary node.
Con - Higher memory consumption.
When changes are made to tables, these changes are recorded in logs. The logs are written to the log buffers in memory. These logs need to be persisted to disk in the log volume. When are log buffers written to the log segments?
1) Log buffers are written when a log buffer becomes full.
2) Log buffers are written to log segments when a commit entry is written.
Why do we use log backups?
1) To roll forward committed transactions to ensure no data loss since the last successful backup.
2) To free up space in the file system that contains the log volumes.
What is required to recover a database to a specific point in time?
latest full data backup before the specified point in time +
the latest differential data backup +
the all incremental data backups since that differential data backup +
the log entries from the log backups between the latest used data backup and the specified point in time
What is RTO and RPO in the context of high availability/disaster recovery?
RTO defines how long it takes to start the recovery from discovery
RPO defines how much data loss is acceptable
What are the two types of delta backups called? Describe them.
Incremental Backup: Stores the data changed since the last full or last delta backup
Differential Backup – stores all data changed since the last full backup.
If the indexserver in a worker node becomes unresponsive. What happens next in the context of auto service restart?
The daemon will try to restart the index server on that node as per auto service restart
How do backups to third party backup servers take place?
3rd party backup agent runs on the HANA server, implements backing for HANA API. It communicates with 3rd party backup server. Backups transferred via pipe.
What is data tiering?
Assigning data to various storage areas and storage media often by their temperature classification.
Which service acts as an SQL port to allow connection to the system DB?
Nameserver.
In a multitenant environment, the system DB does not have its own _______?
Indexserver
What does the max_concurrency parameter control?
Limits the number of concurrently running threads used by the job executor
In the context of cross-database queries, what does a user (U2) that exists in database DB2 need to have done to be able to access objects in another tenant database, DB1?
A user, U1 needs to be created in DB1 with user, U2 in DB2 as a remote identity
What are extension nodes used for?
Extension nodes allow storage of warm data greater than the RAM available
How can we access data in a remote source that needs to be queried in a HANA database, when the data is not in a HANA system?
By using Smart Data Access (SDA) with virtual tables.
Enterprise Information Management is made of two components, what are they? Describe them.
Smart Data Quality/SDQ (Provides functions for improving the data quality) and Smart Data Integration/SDI (Provides functions for acquiring data and integrating the data)
True or False: Data services is a built-in HANA component.
False
Where does the SAP system landscape transformation (SLT) server create logging tables?
In the source system
Where does the SAP system landscape transformation (SLT) server carry out data transformations?
On the SLT system
In the context of SAP system landscape transformation (SLT) server, what connection types exist between the SLT, source, and HANA systems?
DB connection between the SLT system and HANA system and RFC connection between the source system and SLT system
