Creating AWS RDS Oracle instances for personal dev projects

AWS (as in any cloud provider) for spinning something up ondemand for a quick test. Need an Oracle database? Couple of clicks and you’re done.

With the RDS options on AWS be careful of the EC2 type if you’re trying to keep costs down. For example if you leave the DB Instance Size at ‘Standard’, you only get db.m4 and db.m5 instances which are definitely not your cheapest options:

If you select ‘Burstable’, the smallest instance size you can select is a db.t3.small:

According to the Simple Monthly Calculator, at 100% 24/7 usage for a month the cost is around $57/month:

If you’re usage for development is going to be low, and/or if you can stop the instance when you’re not using it (and only pay for storage), for example at 10% usage for a month you can get the cost down to around $8:

Creating an AWS Aurora Serverless database

I’m looking for a low cost managed db in the cloud for a small project, so thought I’d take a look at setting up an Aurora Serverless db, as depending on usage (and my usage will be very low) it looks like it’s definitely the cheapest of all AWS RDS options.

From the Console, from RDS, I pressed the ‘Create Database’ button:

If you select Aurora, the Serverless option is way down the page here:

I kept all the defaults, but changes the capacity to the smallest options:

After taking note of generated credentials and pressing the last ‘Create Database’ button, the dialog said it would take a couple of minutes to provision, and it sure did. I wasn’t timing it but it was at least 10 minutes before it was ready. This was probably the longest I’ve every waiting to provision anything on AWS.

Once it was ready, I tried to use the online query editor, but looks like there’s an additional step to create a user:

This option is under Network and Security:

After applying the change with the immediate option, I created a test table:

Inserted a row:

And then selected all rows:


Looks good so far!

MacOS Catalina, npm global installs and zsh

As MacOS switched to zsh replacing bash a while back (you may have noticed the prompt to change to zsh in your Terminal), I keep coming across a few issues that I need to work around. My latest was that I noticed apps I’d installed globally with ‘npm install -g’ were no longer in my path.

Following a combination of suggestions from answers to this question, I added the following line to ~/.zshrc to add the npm global install dir to my path:

export PATH="$PATH:$HOME/.npm-global/bin"

Kafka producer: ip addresses and host names

Scenario:

  • Kafka client 2.4.0
  • Java 1.8.0_151
  • Kafka cluster is running on a machine with hostname ‘ubuntu-confluent’
  • Producer has bootstrap.servers=10.0.0.x (ip of same host as ubuntu-confluent)
  • At runtime, it appears hosthame is passed back to client
  • Subsequent network calls from client back to cluster appear to use hostname instead of ip, and it fails

Exception on client:

2020-03-08 21:26:06,144 [kafka-producer-network-thread | producer-1] WARN org.apache.kafka.clients.NetworkClient [] - [Producer clientId=producer-1] Error connecting to node ubuntu-confluent:9092 (id: 0 rack: null)
java.net.UnknownHostException: ubuntu-confluent: nodename nor servname provided, or not known
at java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method) ~[?:1.8.0_151]
at java.net.InetAddress$2.lookupAllHostAddr(InetAddress.java:928) ~[?:1.8.0_151]
at java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1323) ~[?:1.8.0_151]
at java.net.InetAddress.getAllByName0(InetAddress.java:1276) ~[?:1.8.0_151]
at java.net.InetAddress.getAllByName(InetAddress.java:1192) ~[?:1.8.0_151]
at java.net.InetAddress.getAllByName(InetAddress.java:1126) ~[?:1.8.0_151]
at org.apache.kafka.clients.ClientUtils.resolve(ClientUtils.java:104) ~[kafka-clients-2.4.0.jar:?]

Easiest fix is to just add an entry to /etc/hosts.