AWS Certified SAP on AWS Specialty PAS-C01 (PAS-C01) — Questions 76–150

Option A is correct because EBS Snapshots provide point-in-time backups and can be taken while the volume is attached and in use (with some performance impact). Option B is wrong because AWS Backup can orchestrate snapshots but is not the service that directly creates them. Option C is wrong because S3 is object storage, not for EBS backups.

Option D is wrong because Storage Gateway is for hybrid storage, not for backing up EBS volumes.

VPC Flow Logs capture IP traffic metadata and can be analyzed to detect packet loss or rejected connections. Option D is correct. Option A is wrong because CloudTrail logs API calls.

Option B is wrong because Direct Connect is for on-premises connectivity. Option C is wrong because VPN is also for external connectivity.

AWS CloudFormation allows infrastructure as code to provision AWS resources in a repeatable manner. Option A is wrong because OpsWorks is for configuration management, not provisioning. Option B is wrong because Elastic Beanstalk is for application deployments, not full infrastructure.

Option D is wrong because CodeDeploy is for application deployment, not infrastructure.

Option B is correct because EBS elastic volumes allow modifying volume type (e.g., to gp3 or io2) and IOPS without detaching the volume or stopping the instance. Option A is wrong because migrating to a larger instance type may improve CPU but not necessarily disk I/O immediately. Option C is wrong because creating a new volume and migrating data requires downtime.

Option D is wrong because adding more volumes with RAID0 requires reconfiguration and possible downtime.

Options A and D are correct because memory usage and CPU usage are critical for HANA performance. Option B is wrong because EBS volume metrics are not HANA-specific. Option C is wrong because SAPS is not a standard CloudWatch metric.

Option E is wrong because network packet loss is not a standard HANA metric.

Option A (Enable CPU hyper-threading) is correct for HANA performance. Option C (Use EBS-optimized instances) is correct for storage performance. Option D (Use placement groups) is correct for high network throughput.

Option B is wrong because swap space is not recommended for HANA. Option E is wrong because T2/T3 instances are burstable and not suitable for production HANA.

Options A, C, and E are correct. A (CPUUtilization) can indicate CPU bottleneck. C (NetworkIn/NetworkOut) shows network throughput.

E (EBS VolumeQueueLength) indicates storage I/O wait. B (StatusCheckFailed) is for instance health, not latency. D (DiskReadBytes) alone is less informative without queue length.

Creating a snapshot of the existing volume is the most efficient way to copy data across Availability Zones. Then, create a new volume in us-east-1b from that snapshot. Detaching the volume and attaching it to an instance in us-east-1b is not possible because the volume is tied to its AZ.

Creating an AMI is unnecessary overhead. Using rsync over the network would require an instance in the same AZ and network bandwidth.

Option A is correct. Using AWS Backup to take frequent snapshots of the database and application volumes and copying them to the DR region can meet the RPO of 1 hour if snapshots are taken every hour. Restoring from snapshots can be automated to meet the RTO.

Option B is wrong because Database Migration Service is for ongoing replication, not backup/restore. Option C is wrong because SAP ASE native replication requires significant changes and licensing. Option D is wrong because Multi-AZ is for high availability within a region, not DR across regions.

Option A is correct because EMRFS with S3 SSE-S3 provides at-rest encryption, and EMR in-transit encryption uses TLS. Option B is incorrect because client-side encryption adds overhead. Option C is incorrect because disabling encryption is not secure.

Option D is incorrect because KMS adds latency.

Option B and Option D are correct: A VPN or Direct Connect (B) provides network connectivity, and EBS snapshots (D) can be used to replicate data if synchronous replication is not feasible. Option A is wrong because S3 Transfer Acceleration is not needed. Option C is wrong because VPC peering does not work across Regions (use Transit Gateway or VPN).

Option E is wrong because AWS Global Accelerator improves latency but is not required.

Option D is correct because EC2 Auto Recovery can automatically recover an instance when system status is impaired. Option A is wrong because the instance is already running. Option B is wrong because rebooting might not fix underlying hardware issues.

Option C is wrong because stopping and starting moves the instance to new hardware, which is similar to recovery, but EC2 Auto Recovery is the automated solution.

Option C is correct because the firewall might block the specific port used by the SOAP service (e.g., 443 or 8080) even if ICMP (ping) is allowed. Option A (DNS) is unlikely if ping works. Option B (SAP router) is not involved.

Option D (SSL certificate) would cause a different error (e.g., certificate validation failure).

Option C is correct: The network ACL is stateless and must allow both inbound and outbound traffic; the outbound rule for ephemeral ports may be missing. Option A is incorrect because the security group allows the port. Option B is incorrect because the NACL allows all inbound traffic.

Option D is incorrect because the security group allows the application server's security group.

SAP HANA on AWS requires /hana/data and /hana/log to be on separate EBS volumes. Using io2 Block Express volumes with RAID 0 for data and RAID 1 for log is recommended. Option A is wrong because gp3 may not provide enough IOPS.

Option B is wrong because /hana/log should be on separate volumes from /hana/data. Option D is wrong because /hana/data should use RAID 0 for performance.

Option B is correct. SAP publishes certified instance types for each SAP application. For SAP BusinessObjects, memory-optimized instances like r5 are certified.

Option A (t3) is burstable and not certified. Option C (c5) is compute-optimized and not certified for BO. Option D (i3) is storage-optimized.

Option B is correct because using EBS io2 Block Express volumes provides consistent low latency and high IOPS for demanding workloads like SAP HANA. Option A is wrong because gp3 volumes have lower performance compared to io2 Block Express. Option C is wrong because increasing instance size may not directly address storage latency.

Option D is wrong because instance store volumes are ephemeral and not suitable for persistent database storage.

Option D is correct because AWS DMS with HANA as target supports minimal downtime using ongoing replication. Option A is incorrect because HSR requires a compatible HANA version setup. Option B is incorrect because DMS cannot replicate from HANA to HANA directly without a source HANA.

Option C is incorrect because AWS Backup cannot replicate data in near-real-time.

Options A, C, and D are correct. Cross-region replication for HANA (A) ensures data consistency. Route53 (C) provides DNS failover.

AWS Backup cross-region copy (D) ensures backups are available. Option B (S3 CRR) is not needed for HANA replication. Option E (AWS Transit Gateway) is for network connectivity, not DR failover.

AWS Key Management Service (KMS) allows creation and management of encryption keys used to encrypt EBS volumes. Option A (CloudHSM) is a hardware security module but more complex. Option C (Secrets Manager) is for secrets.

Option D (ACM) is for certificates.

Option B is correct because placement groups, specifically cluster placement groups, provide low-latency network performance by placing instances in a single Availability Zone with high-bandwidth, low-latency connections. Option A is incorrect because VPC Peering connects networks but doesn't guarantee low latency between instances. Option C is incorrect because AWS Global Accelerator improves global traffic routing but not within a region.

Option D is incorrect because Route 53 is a DNS service.

Option C is correct because EBS volume queue depth can cause I/O delays leading to transaction rollback. Option A is wrong as snapshot creation does not affect live database. Option B is wrong as DB parameter groups don't cause this error directly.

Option D is wrong as Enhanced Networking is for network, not storage.

Options B and D are correct because ensuring correct IAM permissions and bucket policies are common causes of access denied errors. Option A is wrong because KMS key might be separate but not the direct issue. Option C is wrong because S3 Transfer Acceleration is for speed, not access.

Option E is wrong because VPC endpoint is not required if using public S3.

Options A, B, and C are correct. Using security groups, encrypting EBS volumes, and enabling HANA encryption are all security best practices. Option D is incorrect because disabling password policies weakens security.

Option E is incorrect because making the database publicly accessible increases risk.

Options A and C are correct. AWS CloudFormation allows defining infrastructure in templates; AWS Elastic Beanstalk can deploy and manage applications but is less commonly used for SAP. Option B is wrong because OpsWorks is for Chef/Puppet, not typical for SAP.

Option D is wrong because CodeDeploy deploys applications, not infrastructure. Option E is wrong because Service Catalog creates a catalog of products, not direct deployment.

SAP HANA can be deployed as a multi-node cluster using EC2 instances in different Availability Zones with synchronous replication and automatic failover using cluster software like Pacemaker. RDS does not support SAP HANA. A single EC2 instance in one AZ is not highly available.

Multi-AZ RDS is for other databases, not SAP HANA.

Options A and D are correct. CloudWatch can monitor OS and custom HANA metrics. HANA itself provides monitoring via SQL queries (e.g., SAP HANA cockpit).

Option B is incorrect because AWS Config tracks configuration, not performance. Option C is incorrect because VPC Flow Logs monitor network traffic. Option E is incorrect because Trusted Advisor provides best-practice checks, not real-time monitoring.

Option C is correct because HANA scale-out supports rolling updates, allowing nodes to be updated one by one with minimal downtime. Option A is wrong because stopping all services causes full downtime. Option B is wrong because using HSR failover is complex and may cause data inconsistency if not properly managed.

Option D is wrong because creating a new cluster and migrating is not efficient for a minor update.

Using AWS Route 53 health checks to monitor the ASCS instance and a custom script to update DNS records on failover provides a simple and effective HA solution. Option D is correct. Option A is wrong because ALB cannot direct traffic to a single SAP instance for ASCS.

Option B is wrong because Auto Scaling is for scaling out, not HA for stateful services. Option C is wrong because EC2 Auto Recovery recovers the instance but does not handle failover to another AZ.

Option B is correct because the SAP application likely uses the private IP of the old primary; after failover, the new primary has a different IP. Option A is wrong because HANA replication is independent of DNS. Option C is wrong because security groups are not AZ-specific.

Option D is wrong because storage type does not affect connectivity.

Option D is correct because placing application and database in the same placement group with enhanced networking ensures low latency and high throughput. Option A is wrong because different Availability Zones add latency. Option B is wrong because different Regions adds significant latency.

Option C is wrong because S3 is not for database storage.

Option D is correct because network ACLs are stateless and require explicit rules for both inbound and outbound traffic, and ephemeral ports are often missed. Option A is wrong because NACL rules are evaluated before security groups. Option B is wrong because VPC peering is not mentioned.

Option C is wrong because security groups are stateful and would not cause asymmetric traffic issues.

The correct answer is B. The CloudWatch metrics show high read latency (>10 ms) and a queue length peaking at 20, which indicates that the EBS volume is throttling due to insufficient IOPS for the workload. Increasing provisioned IOPS from 5000 to 10000 directly addresses the bottleneck by allowing more I/O operations per second, reducing queue depth and latency.

Since the instance and volumes are in the same Availability Zone, network latency is not a factor, and instance store volumes are ephemeral and unsuitable for persistent SAP HANA data.

Option C is correct because AWS strongly recommends using io2 Block Express volumes for SAP HANA due to their high performance, durability, and ability to meet SAP's high availability requirements. Option A (EFS) is a file system, not block storage. Option B (gp3) is cost-effective but may not provide the consistent low latency needed for HANA.

Option D (Instance Store) provides low latency but is ephemeral and not suitable for persistent data.

Option A is correct because AWS Systems Manager Maintenance Windows allow scheduling of patching with optional approval workflows. Option B is wrong because AWS Config is for compliance, not patching. Option C is wrong because CloudWatch Events can trigger actions but lacks built-in approval gates.

Option D is wrong because CloudFormation is for infrastructure provisioning, not patching.

AWS Backup can automate HANA backups, and Backint agent for SAP HANA is an SAP-certified backup integration. Option A (Amazon S3) is storage, not a backup service. Option C (AWS Storage Gateway) is for hybrid storage.

Option E (AWS DMS) is for migration.

Option B is correct because RDS events and DB instance logs (error log, slow query log) provide information about failover causes. Option A is wrong because CloudTrail logs API calls, not database events. Option C is wrong because VPC Flow Logs are for network traffic.

Option D is wrong because CloudWatch Logs for the OS are not available for RDS.

Option D is correct. To change the instance type, the instance must be in the stopped state. Option A is incorrect because you cannot change instance type while running.

Option B is incorrect because r5.4xlarge and r5.8xlarge are in the same instance family. Option C is incorrect because you can change instance type to any size within the same family without re-attaching volumes.

Option A is correct because read-write SSD tiering reduces latency. Option B is wrong because compression increases CPU overhead. Option C is wrong because it does not affect write latency.

Option D is wrong because it increases latency.

Option D is correct because AWS offers a native HANA multi-AZ solution using HANA System Replication with automatic failover. Option A is wrong because auto-scaling doesn't handle database failover. Option B is wrong because RDS custom does not support HANA.

Option C is wrong because manual DNS change is not automatic.

AWS Snowball Edge can be used with SAP DMO to transfer data offline, especially for large databases. AWS Direct Connect improves network speed but is not a service for acceleration. AWS DataSync is for file data.

AWS DMS is not used with DMO.

Option A is correct because SAP Central Services can be deployed in an ASCS/ERS cluster using two EC2 instances with shared storage (like EBS Multi-Attach or FSx for Windows) and a floating IP. Option B (RDS for Oracle) can be HA but not for SCS. Option C (single instance with EBS snapshots) does not provide HA.

Option D (Application Load Balancer) is for HTTP traffic, not SAP protocols.

Options B and D are correct. Option B creates a standby instance in another AZ using replication. Option D creates automated snapshots for point-in-time recovery.

Option A is wrong because increasing instance size does not improve availability. Option C is wrong because Multi-AZ RDS is for RDS, not EC2-based HANA. Option E is wrong because Amazon FSx does not replace the database.

Option A is correct because RDS for SAP HANA uses a database port (e.g., 3xx17) but also requires an additional port for internal services. If the security group does not allow traffic on that additional port, connections may fail. Option B is incorrect because NACLs are stateless but properly configured.

Option C is incorrect because RDS does not use DNS-based routing issues typically. Option D is incorrect because the issue is not about routing tables if within same VPC.

Options A, B, and D are correct. Pre-provisioning EBS volumes (A), AMIs (B), and Elastic IPs (D) helps reduce RTO. Option C is wrong because EC2 instances should not be running to avoid costs.

Option E is wrong because Amazon RDS for SAP HANA is not used; HANA runs on EC2.

Options A and C are correct. HANA System Replication across AZs (A) and using EBS io2 volumes (C) are key for HA. Option B is wrong because EFS is not used for database data.

Option D is wrong because Auto Scaling is not for HANA. Option E is wrong because S3 Standard-IA is for backups, not primary storage.

Options A, B, and D are correct. CloudWatch metrics help identify resource bottlenecks. ALB access logs show request-level details.

Detailed monitoring provides high-resolution metrics. Option C (VPC Flow Logs) helps with network analysis but not application latency. Option E (AWS Config) tracks configuration changes, not performance.

Option B is correct: for HSR scale-out, a multi-AZ cluster with automatic failover provides high availability. Option A is incorrect because a single AZ is a single point of failure. Option C is incorrect because HSR active/active is not supported for scale-out.

Option D is incorrect because Multi-AZ is for RDS, not HANA on EC2.

Options B and C are correct. SAP HANA System Replication is the standard method for replicating data between HANA instances across regions. Amazon EBS Snapshots can be copied to another region and used to restore volumes.

Option A (EFS) is not suitable for HANA data volumes. Option D (S3 Transfer Acceleration) is not a replication method for HANA. Option E (DMS) is for database migrations, not ongoing replication.

Adding more SAP application servers (ASCS) behind an Application Load Balancer distributes load and provides HA. Option A is wrong because vertical scaling is limited. Option B is wrong because ASG alone does not register with SAP load balancer.

Option C is wrong because launching manually is not automated.

Options A and D are correct. Synchronous replication provides near-zero RPO but increases latency; asynchronous mode with a short RPO can meet <15 minutes with proper network. Using a multi-AZ deployment ensures rapid failover for RTO <2 hours.

Option B (single AZ) increases risk. Option C (separate region) increases latency. Option E (EBS snapshots) does not meet RPO.

Option A is correct because HANA system replication with synchronous mode provides near-zero RPO. Option B is wrong because asynchronous replication may have higher RPO. Option C is wrong because EBS snapshots have higher RPO.

Option D is wrong because restoring from S3 is slow and has higher RPO.

Option B is correct because cross-region snapshot copies are asynchronous and may fail if the source region becomes unavailable. Option A is wrong because IAM does not prevent copying. Option C is wrong because KMS keys can be used cross-region with proper configuration.

Option D is wrong because the snapshot ID would not become invalid.

Option C is correct because stale file locks can be released by enabling automatic file lock duration or using the FSx file lock command. Option A is wrong because S3 does not support file locking. Option B is wrong because EFS supports NFSv4 locks but this is Windows.

Option D is wrong because increasing throughput does not address lock issues.

For large data volumes with limited bandwidth, using AWS Snowball devices for offline transfer is most efficient. After the initial load, AWS DMS can replicate ongoing changes. Option A is correct.

Option B (AWS Direct Connect) would require high bandwidth and time. Option C (VPN + DMS) is slow for initial load. Option D (AWS DataSync) is also network-based.