Hi Team, I am in process of building singer based ...
# ask-community-for-troubleshooting
r
Hi Team, I am in process of building singer based connector. I am trying to run below command ./gradlew airbyte integrationsconnectorssource oracle py singerbuild and getting the below error. Plugin [id: 'com.gradle.enterprise', version: '3.4.1'] was not found in any of the following sources: - Gradle Core Plugins (plugin is not in 'org.gradle' namespace) - Plugin Repositories (could not resolve plugin artifact 'com.gradle.enterprisecom.gradle.enterprise.gradle.plugin3.4.1') Searched in the following repositories: Gradle Central Plugin Repository
m
What OS are you? any chances to move the singer code to Airbyte CDK? 😛
r
Using Ubuntu on wsl, I am trying to build oracle db based connector based on pipelinewise-tap-oracle.
m
The Source Oracle of Airbyte doesn’t attempt what you need?
r
it does work, but performance is not good. 20 million records table took 15 hr. table size has 300 million records. Hoping singer based connector might give good performance. Apart from this, wanted to add start_date feature (means instead of full load, will load via state file) and batch size. Airbyte's oracle connector does not have any feature for modify read batch_size, default batch size 1000.
m
@Rahul Patel latest version of connector should use dynamic batch fetching
about the problem: running the python commands did it work?
r
yes, python cmd works fine, read, discover tables works fine. Dynamic batch fetching is calculating max 35k records in a batch based on row size which still gives low performance for same table.
I did resolve above by adding cacert of site in java11 cacerts. but same error is starting to throw all dependencies.
I had to upgrade java 8 to 11 (since add-exports parameter is not supported) then added cacerts
m
Airbybte uses Java 17
r
does it means that for successful gradle build, I need java 17 ? @Marcos Marx (Airbyte)
m
Probably helps use the right version of the dev stack 😅
😅 1