Change Log
New Features
UI Support
- Seamlessly navigate and interact with the platform through an enhanced user interface, boosting user satisfaction and productivity.
Workspace
- Empower users to curate personalized workspaces with jobs and notebooks, enhancing collaboration and efficiency in project management.
Spark SQL
- Execute queries and leverage Spark compute seamlessly with our latest update, featuring Spark SQL support within jobs.
Notebook
- Dynamically execute Scala and Python notebooks to cater to varying workload demands, enhancing data analysis flexibility.
Cluster-Level Global Configuration
- Optimize Spark clusters by configuring settings at the cluster level, improving performance and resource utilization.
Access Management
- Fine-tune workspace accessibility and permissions with distinct roles: "Can View," "Can Run," "Can Edit," and "Can Manage," ensuring streamlined collaboration.
Kubernetes Support
- Integrate Kubernetes architecture seamlessly with Yeedu's core services, ensuring scalability and fault tolerance for enterprise-grade deployments.
Multicloud/Multi-zone Capability
- Simplify cluster management with a unified Cloud Environment, consolidating cloud-related configurations and reducing administrative overhead.
Increased Reliability
- Notebook Idle Timeout: Optimize resource usage by automatically terminating idle notebooks after 15 minutes, ensuring efficient resource allocation.
Vault Support
- Securely store credentials with vault support, enhancing protection of sensitive information.
Tab Complete Support for CLI
- Enhance navigation and productivity within the CLI interface with effortless command auto-completion triggered by the Tab key press.
Minor Changes
- Nginx Configuration: Updated to utilize upstream servers.
- Kubernetes Ingress: Implemented the
setup-ingress.sh
script for configuration. - Randomized Kubernetes Passwords: Replaced default passwords with randomly generated ones.
- Telegraf Configuration: Included network interfaces in the configuration file.
- Checkpoint Volume Mounts: Introduced volume mounts for the checkpoint file in the
yeed-reactors-logs-collector
service. - Kubernetes Helm: Segregated third-party and core services.
- Cluster Resource Checks: Added checks for Kubernetes cluster resources.
- GPU Drivers: Updated GPU drivers setup.
- Enhanced Kubernetes Functionality:
- Conditional
storageClassName
in PVC creation. - Added labels.
- Implemented HA for MQ, Redis, and Vault.
- Applied node affinity and anti-affinity.
- Set limits for Pods.
- Added support for helm upgrades.
- Added nginx Ingress annotations for improved proxy settings and timeouts.
- Configured Grafana-admin username and password from environment variables.
- Conditional
- History Server Improvements: Various enhancements made.
- Telegraf Metrics: Modified Telegraf to send metrics to the staging table.
- Password Handling: Added support for special characters in passwords.
- SSL Support: Added SSL support for history server nginx.
- History Server Configurations: Added configurations in the nginx conf file.
- Spark Scratch Configurations: Incorporated additional configurations.
- Azure Custom Script Extension: Removed the custom_script extension.
- Setup Custom Spark Configs: Introduced setup_custom_spark_configs with redaction of secrets.
- Azure CFE Zone and Region: Adjusted according to data model changes.
- OSM Container Loggers: Implemented loggers.
- Terraform Upgrade: Upgraded Terraform to version 1.7.4.
- Scratch Disk Setup: Modified to use RAID.
- Password Security: Handled special characters in passwords.
- CUDA Support Column: Added support for the new "has_cuda_support" column.
- Linux Distro Validation: Implemented validation.
- GCP Disk Change: Changed local SSD disk from NVME to SCSI.
- Docker Image Update: Updated the spark docker image name accordingly.
- Logging Methods: Enhanced logging methods.
- CLI Installation Verification: Verified installations.
- Local Storage Mounting: Updated
cfe.py
for mounting local storage to/scratch
. - GCP Terraform GPU Changes: Made necessary adjustments.
- Credential Security: Ensured credentials are hidden in logs.
- Telegraf Network Configuration: Added
--network=host
to Telegraf. - Spark Environment: Added
SPARK_LOCAL_DIRS
inspark-env.sh
for the worker container in standalone mode.
Bug Fixes
- User Input Cleanup: Addressed clean-up issue.
- Grafana and MQ Ports: Fixed port issues in Docker.
- SSL Connection: Resolved SSL connection issue in
Validate_third_party_services
. - Grafana CPU Usage: Fixed CPU usage issue.
- Permission Denied Issues: Resolved for
cfe.py
and updated configurations. - Hive Metastore: Fixed metastore issue.
- Spark Master Container: Addressed start issues.
- RHEL Images Space: Fixed space issues.
- Bootstrap Shell Script: Corrected query.
- Reactor Container: Fixed startup command issues.
- Setup Reactors: Addressed spin issue for Spark version 2.4.8.
- Unsupported Notebook Versions: Fixed setup reactors.
- Ubuntu CUDA Version: Addressed version issues.
- Azure Mounted Disks: Fixed
/mnt
mounted disks issue. - AWS Worker Numbers: Fixed standalone worker numbers.