High Availability and Business Continuity 15% Flashcards
DR Scenario that uses S3, Import/Export, Storage GW
Back and Restore
DR scenario that has minimal version of infra. EC2, RDS, AMIs, CF, pre-allocated ENIs
Pilot Light
DR scenario that reduces recovery time by having some services already ready, horizontal scaling with elbs, manual dns or auto route 53 dns based on health check, and use of autoscaling to accommodate additional load
Warm Standby
DR scenario that is active/active between on prem and AWS. Route53 can send traffic to either or both, autoscaling to handle additional load
Mutlisite
Minimum time to recovery from Glacier
3 hours
Does DynamoDB over Cross-Region replication
Yes
This service has the ability to replicate snapshots or run read replicas in other regions
RDS
The data storage service allows you to take snapshots to S3 and replicate them to another region
Redshift
This database service uses Always on Availability Groups, and Mirroring for HA
SQL
DB service that uses async replication to other regions for ha
MySQL
DB services that uses DataGuard and RAC for HA
Oracle
Situations that can trigger multi-az automatic failover
Loss of connectivity to primary db, storage or host failure, patching, or manually with reboot of primary db
These multi-az dbs use amazon’s az failover tech
Oracle, PostgreSQL, MySQL, and MariaDB
This DB uses it’s own mirroring for failover within an az
SQL
This DB uses copies of the data in a db cluster across multiple AZ’s
Aurora
How many Read Replicas can you have per a single instance
5
Uses for Read Replicas
offloading read traffic, continual reads even during backed up or maintained, business reporting, getting io closer to customers, dr capabilities, data center migrations
MySQL 5.6 uses which engines? Which one is supported by AWS?
InnoDB or MYisam. InnoDB is only supported by AWS
Which versions of PostgreSQL is supported
9.3.5 or newer
Which versions of MariaDB is supported
All current versions
Which versions of MySQL and Oracle are supported
All current versions
T/F. When creating RR of your RDS database with out Multi-AZ enable you will not experience any outage.
False. Only when you have multi-az enable will you be able to take a RR without interruption. Can go offline for about a minute during snapshotting
Which databases support RR in a different region?
MySQL, PostgeSQL, MariaDB (RDS), and also Aurora
Which databases do not support RR’s in a different region?
SQL and Oracle
T/F. Read replicas can also be mutli-az enabled
False
Which DB services allows for RR’s of RR’s
MySQL. Can increase latency
T/F. If creating a replica of an encrypted database the target can be un-encrypted
False. The encryption must match
T/F. DB snapshots and backups can be taken from RR’s
False. Only can be done on the primary db
What DB should you use if the application does not require ACID, joins or SQL support?
DynamoDB
Gateway cached volumes size and totals supported
Each volume can be 32TB with 32 volumes supported. Total size of 1PB
Gateway stored volumes size and total supported
Each volume can be 16TB with 32 volumes support. Total storage 512TB.
Storage amounts supported by Gateway VTL
1500 virtual tapes up to 1PB but can be archived to Glacier for unlimited storage
AWS service that can allow you to have unlimited storage and r/w files via NFS
File Gateway
Largest file size supported by S3 and File Gateway
5TB
Storage Gateway facts
Can be deployed on premise on VMWare or HyperV, or on an EC2 instance. Can be used with Direct Connect. Can be throttled
Storage Gateway hardware requirement
4 or 8 CPUs, 16GB RAM, 80 local disk
Storage requirements for Gateway Cached data storage
For 200GB volume you should use ~300GB so there is an upload buffer
Storage requirements for Gateway Stored data storage
1:1 ratio since the low latency storage is completely stored on premise and also replicated to AWS
What ports are used by Storage Gateway?
80 (only for activation), 443, 3260 (iSCSI), UDP 53
T/F EBS volumes for instances can be created from Storage Gateway snapshots?
True.
Storage gateway snapshot are stored where?
S3
How are Storage Gateway snapshots run?
Scheduled or ad-hoc
Access times to virtual tapes?
Start instantly
Access times to restore from virtual tape shelf
around 24 hours
How is the data encrypted for Storage Gateways
Encrypted in transit using SSL and at rest using AES256
Snowball capacities
Standard is 80TB and Edge has 100TB
T/F Snowball Edge can run Lambda function
True. Edge has onboard compute
Snowball can import and export to what service(s)
S3. Glacer can also be used but an export must be done first to write to the Snowball
T/F AWS Export can export previous versions of data
False. Only current versions supported
When should you consider using Snowball?
When you data transfer can take more than a week to complete
T/F Using Direct Connect is a cheaper option to move large amounts in a one-time transfer of data into AWS than Snowball
False
T/F Using Snowball you can import directly to EBS.
False. This can be done however using the Import/Export Service
Services that have automated backups
RDS, Elasticache (Redis Only), Redshift
T/F. When deleting an RDS instance all backups will be deleted including the manually initiated onces
False, only the automatic backups will be deleted. Manually backed up dbs will need to be deleted.
How to change the engine type on a database such as SQL Standard to Enterprise
This can be done by dong a restore to a new instance.
T/F. Elasticache Backups are available on Redis and Memcached
False. Backups are only available on Redis
Elasticache backups performance considerations
Performance degraded during backup and is done on the entire cluster so should only be done after hours
Redshift backup schedule and storage
Automated by default with a 1 day retention, stored in S3 and only changed data stored