Streaming from Oracle Tables to Redshift [closed] - amazon-s3

Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 4 years ago.
Improve this question
I am new to Redshift and AWS eco-system. I am looking for options or best practices for streaming data changes from Oracle EBS tables on premise to Redshift.
Should S3 be used as the staging area? Thus Oracle->S3->Redshift? Is that good for real-time streaming Oracle tables to Redshift?
Any way to bypass S3 staging and do Oracle->AWS Kinesis(Firehose)->Redshift? If so, are there such scenarios I can read-up on?
What about using Kafka instead of AWS Kinesis?
Can AWS Kinesis or Kafka pull directly from an on-premise Oracle instance?
Are there other alternatives/components, ETL tools for near real-time or almost real-time data load to Redshift?
There is a large number of tables to stream from Oracle, which is on-prem. I am new to Redshift but familiar with Oracle, SQL Server, PG. Sorry if I am totally of beat here.
Please help :) Any thoughts and/or references would be highly appreciated...

As per docs here, 1 and 2 are the same, essentially. You won't bypass S3 by using firehose, just mask it. And Firehose is currently useless if you have lots of tables on more than one cluster. Unless of course you plan to automate the process of sending support requests to AWS Support to increase limits (I was thinking about it, don't laugh).
I'd go for loading using COPY command from S3.
Inserts are currently slow, and I mean SLOW. Do not use methods that would generate insert statements under the hood.
My use cases:
Apache Storm streaming events to Redshift, using S3 as staging area. Works fine for hundreds of thousands of events per table per day, dozens of tables per database, several databases per cluster, couple of clusters. We use API Gateway, AWS Lambda and S3 as staging area for second process. Works just as well for tens of thousands of events per day, couple of different clusters, several databases on each cluster, one table is loaded this way in each database.
You can, in theory, issue COPY command using SSH, but then you have to prepare manifest files on (wait for it) ... S3. So I have no idea why you wouldn't use S3 for staging data storage anyway.
As for streaming data from on-premises Oracle to S3, it's a whole different topic entirely, and you should look for answers from someone proficient with Oracle. I'd look at CDC, but I'm not an Oracle pro, so can't tell if this is a good approach.
I hope this helps.

Related

Replicate data from cloud SQL postgres to bigQuery

I am looking for the recommended way of streaming database change from cloud SQL (postgres) to bigQuery ? I am seeing that CDC streaming does not seems available for postgres, does anyone know the timeline of this feature ?
Thanks a lot for you help.
Jonathan.
With Datastream for BigQuery, you can now replicate data and schema updates from operational databases directly into BigQuery.
Datastream reads and delivers every change—insert, update, and delete—from your MySQL, PostgreSQL, AlloyDB, and Oracle databases into BigQuery with minimal latency. The source database can be hosted on-premises, on Google Cloud services such as Cloud SQL or Bare Metal Solution for Oracle, or anywhere else on any cloud.
https://cloud.google.com/datastream-for-bigquery
You have to create an ETL process. That will allow you to automatically transform data from Postgres into BigQuery. You can do that using many ways, but I will point you to the two main approaches that I've already implemented:
Way 1:
Set Up the ETL Process manually:
Create your ETL using open source tools...
This method involves the use of the COPY command to migrate data from PostgreSQL tables and standard file-system files. It can be used as a normal SQL statement with SQL functions or PL/pgSQL procedures which gives a lot of flexibility to extract data as a full dump or incrementally. You need to know that it is a time-consuming process and would need you to invest in engineering bandwidth!
Also, you could try different tech stacks to implement the above, and I recommended this one Java Spring Data Flow
Way 2:
Using DataFlow
You can automate the ETL process using GCP's DataFlow without coding your own solution. It is faster and it cost, of course.
DataFlow is Unified stream and batch data processing that's
serverless, fast, and cost-effective.
Check more details and learn in a minute here
Also check this

How to transform data from S3 bucket before writing to Redshift DW?

I'm creating a (modern) data warehouse in redshift. All of our infrastructure is hosted at Amazon. So far, I have setup DMS to ingest data (including changed data) from some tables of our business database (SQL Server on EC2, not RDS) and store it directly to S3.
Now I must transform and enrich this data from the S3 before I can write it to Redshift. Our DW have some tables for facts and dimensions (star schema), so, imagine a Customer dimension, it should contain not only the customer basic info, but address info, city, state, etc. This data is spread amongst a few tables in our business database.
So here's my problem, I don't have a clear idea of how to query the S3 staging area in order to join these tables and write it to my redshift DW. I want to do it using AWS services like Glue, Kinesis, etc. i.e. full serverless.
Can Kinesis accomplish this task? Would it make things easier if I moved my staging area from S3 to Redshift since all of our data is highly relational in nature anyway? If so, the question remains, how to transform/enrich data before saving it on our DW schemas? I have searched everywhere for this particular topic but information on it is scarse.
Any help is appreciated.
There are a lot of ways to do this but one idea is to query the data using Redshift Spectrum. Spectrum is a way to query S3 (called an external database) using your Redshift cluster.
Really high-level, one way to do this would be to create a Glue Crawler job to crawl your S3 bucket, which creates the External Database that Redshift Spectrum can query.
This way, you don't need to move your data into Redshift itself. Likely, you will want to keep your "staging" area in S3 and only bring into Redshift the data that is ready to be used for reporting or analytics, which would be your Customer Dim table.
Here is the documentation to do this: https://docs.aws.amazon.com/redshift/latest/dg/c-getting-started-using-spectrum.html
To schedule the ETL SQL: I don't believe there is a scheduling tool built into Redshift but you can do that in a few ways:
1) Get an ETL tool or set up CRON jobs on a server or Glue that schedules SQL scripts to be ran. I do this with a Python script that connects to the database then runs the SQL text. This would be a little bit more of a bulk operation. You can also do this in a Lambda function and have it be scheduled on a Cloudwatch trigger which can be on a cron schedule
2) Use a Lambda function that runs the SQL script that you want that triggers for S3 PUTs into that bucket. That way the script will run right when the file drops. This would be basically a realtime operation. DMS drops files very quickly so you will have files dropping multiple times per minute so that might be more difficult to handle.
One option is to load the 'raw' data into Redshift as 'staging' tables. Then, run SQL commands to manipulate the data (JOINs, etc) into the desired format.
Finally, copy the resulting data into the 'public' tables that users query.
This is a normal Extract-Load-Transform process (slightly different to ETL) that uses the capabilities of Redshift to do the transform.

Dump materialize aggregation from BigQuery to SQL server, Dataflow vs Airflow

I use a BigQuery dataset as data lake to store all records/events level data, and a SQL server to store aggregated reports that are updated regularly. Because the reports will be accessed frequently by clients via web interface, and each report aggregates large amount of data, so storing it BigQuery is a no go.
What is the best practise for doing this? Internally we have 2 ideas running around:
Run a Dataflow batched job every X hr to recalculate the aggregation and update the SQL server. It will need a scheduler to trigger the job, and the same job can be used to backfill all data.
Run an Airflow job that does the same thing. A separate job will be needed for backfill (but can still share most of the code with the regular job)
I know Dataflow does well in terms of processing chunks of data in parallel, but I wonder about Airflow's performance, as well as the risk of exhausting connection limit
Please check this answer from a previous similar question
In conclusion: Using Airflow will result in a more efficient way to manage all the process from the workflow. A solution that Google offers based on Airflow is Cloud Composer.

Managing data in two relational databases in a single location

Background: We currently have our data split between two relational databases (Oracle and Postgres). There is a need to run ad-hoc queries that involve tables in both databases. Currently we are doing this in one of two ways:
ETL from one database to another. This requires a lot of developer
time.
Oracle foreign data wrapper on our
Postgres server. This is working, but the queries run extremely
slowly.
We already use Google Cloud Platform (for the project that uses the Postgres server). We are familiar with Google BigQuery (BQ).
What we want to do:
We want most of our tables from both these databases (as-is) available at a single location, so querying them is easy and fast. We are thinking of copying over the data from both DB servers into BQ, without doing any transformations.
It looks like we need to take full dumps of our tables on a periodic basis (daily) and update BQ since BQ is append-only. The recent availability of DML in BQ seems to be very limited.
We are aware that loading the tables as is to BQ is not an optimal solution and we need to denormalize for efficiency, but this is a problem we have to solve after analyzing the feasibility.
My question is whether BQ is a good solution for us, and if yes, how to efficiently keep BQ in sync with our DB data, or whether we should look at something else (like say, Redshift)?
WePay has been publishing a series of articles on how they solve these problems. Check out https://wecode.wepay.com/posts/streaming-databases-in-realtime-with-mysql-debezium-kafka.
To keep everything synchronized they:
The flow of data starts with each microservice’s MySQL database. These
databases run in Google Cloud as CloudSQL MySQL instances with GTIDs
enabled. We’ve set up a downstream MySQL cluster specifically for
Debezium. Each CloudSQL instance replicates its data into the Debezium
cluster, which consists of two MySQL machines: a primary (active)
server and secondary (passive) server. This single Debezium cluster is
an operational trick to make it easier for us to operate Debezium.
Rather than having Debezium connect to dozens of microservice
databases directly, we can connect to just a single database. This
also isolates Debezium from impacting the production OLTP workload
that the master CloudSQL instances are handling.
And then:
The Debezium connectors feed the MySQL messages into Kafka (and add
their schemas to the Confluent schema registry), where downstream
systems can consume them. We use our Kafka connect BigQuery connector
to load the MySQL data into BigQuery using BigQuery’s streaming API.
This gives us a data warehouse in BigQuery that is usually less than
30 seconds behind the data that’s in production. Other microservices,
stream processors, and data infrastructure consume the feeds as well.

Tool to "Data Load" or "ETL" -- from SQL Server into Amazon Redshift

I am trying to figure out decent but simple tool which I can host myself in AWS EC2, which will allow me to pull data out of SQL Server 2005 and push to Amazon Redshift.
I basically have a view in SQL Server on which I am doing SELECT * and I need just put all this data into Redshift. The biggest concern is that there is a lot of data, and this will need to be configurable so I can queue it, run as a nighly/continuous job, etc.
Any suggestions?
alexeypro,
dump tables to files, then you have two fundamental challenges to solve:
Transporting data to Amazon
Loading data to Redshift tables.
Amazon S3 will help you with both:
S3 supports fast upload of files to Amazon from your SQL server location. See this great article. It is from 2011 but I did some testing a few months back and saw very similar results. I was testing with gigabytes of data and 16 uploader threads were ok, as I'm not on backbone. Key thing to remember is that compression and parallel upload are your friends to cut down the time for upload.
Once data are on S3, Redshift supports high-performance parallel load from files on S3 to table(s) via COPY SQL command. To get fastest load performance pre-partition your data based on table distribution key and and pre-sort it to avoid expensive vacuums. All is well documented in Amazon's best practices. I have to say these guys know how to make things neat & simple, so just follow the steps.
If you are coder you can orchestrate the whole process remotely using scripts in whatever shell/language you want. You'll need tools/libraries for parallel HTTP upload to S3 and command line access to Redshift (psql) to launch the COPY command.
Another options is Java, there are libraries for S3 upload and JDBC access to Redshift.
As other posters suggest, you could probably use SSIS (or essentially any other ETL tool) as well. I was testing with CloverETL. Took care of automating the process as well as partitioning/presorting the files for load.
Now Microsoft released SSIS Powerpack, so you can do it natively.
SSIS Amazon Redshift Data Transfer Task
Very fast bulk copy from on-premises data to Amazon Redshift in few clicks
Load data to Amazon Redshift from traditional DB engines like SQL Server, Oracle, MySQL, DB2
Load data to Amazon Redshift from Flat Files
Automatic file archiving support
Automatic file compression support to reduce bandwidth and cost
Rich error handling and logging support to troubleshoot Redshift Datawarehouse loading issues
Support for SQL Server 2005, 2008, 2012, 2014 (32 bit and 64 bit)
Why SSIS PowerPack?
High performance suite of Custom SSIS tasks, transforms and adapters
With existing ETL tools, an alternate option to avoid staging data in Amazon (S3/Dynamo) is to use the commercial DataDirect Amazon Redshift Driver which supports a high performance load over the wire without additional dependencies to stage data.
https://blogs.datadirect.com/2014/10/recap-amazon-redshift-salesforce-data-integration-oow14.html
For getting data into Amazon Redshift, I made DataDuck http://dataducketl.com/
It's like Ruby on Rails but for building ETLs.
To give you an idea of how easy it is to set up, here's how you get your data into Redshift.
Add gem 'dataduck' to your Gemfile.
Run bundle install
Run datatduck quickstart and follow the instructions
This will autogenerate files representing the tables and columns you want to migrate to the data warehouse. You can modify these to customize it, e.g. remove or transform some of the columns.
Commit this code to your own ETL project repository
Git pull the code on your EC2 server
Run dataduck etl all on a cron job, from the EC2 server, to transfer all the tables into Amazon Redshift
Why not Python+boto+psycopg2 script?
It will run on EC2 Windows or Linux instance.
If it's OS Windows you could:
Extract data from SQL Server( using sqlcmd.exe)
Compress it (using gzip.GzipFile).
Multipart upload it to S3 (using boto)
Append it to Amazon Redshit table (using psycopg2).
Similarly, it worked for me when I wrote Oracle-To-Redshift-Data-Loader