Looks like Maven repo is behind search.maven.or, instead of central.maven.org. i cant reach "central" directly.

The documentation should reflect as wget http://search.maven.org/maven2/org/apache/kudu/kudu-spark2_2.11/1.9.0/kudu-spark2_2.11-1.9.0.jar instead of wget http://central.maven.org/maven2/org/apache/kudu/kudu-spark2_2.11/1.9.0/kudu-spark2_2.11-1.9.0.jar

It is crashing out and giving me:

FAILED!!! (signal: 0, exit code: 1)

Running multiple launches concurrently can cause some of them to fail.
The launch process creates and deletes *.signed and *.urls files under setup/terraform/resources and if one launch process deletes one of those files while the terraform of another launch is uploading the directory to its cluster instance, the latter terraform process will fail due to trying to copy a file that no longer exists.

Calls to http://ifconfig.me can fail with transient errors and some parts of the code do not retry the operation, causing the deployment to fail.

As the exercise was being followed-along, this graph did apply to make the NiFi Flow work. We had to remove the "Impala::" piece from the name to make it work. This had to be worked around and it worked finally without it. Please add a note or comment in case other users find the same problem.

Changes in the CDSW web pages caused API authentication with VizApps to stop working.
The automation script fail to authenticate correctly with VizApps and do not successfully set up the VizApps credentials.

this is leading to:

Error: "/32" is not a valid CIDR block: invalid CIDR address: /32

  on security_group.tf line 21, in resource "aws_security_group" "workshop_web_sg":
  21: resource "aws_security_group" "workshop_web_sg" {



Error: "/32" is not a valid CIDR block: invalid CIDR address: /32

  on security_group.tf line 57, in resource "aws_security_group_rule" "workshop_ssh_sg_rule":
  57: resource "aws_security_group_rule" "workshop_ssh_sg_rule" {

Cloudera Manager 7.11.3.2+ Agent on ECS 1.5.2 instance will fail to start if python3.8 is not installed
https://docs.cloudera.com/cdp-private-cloud-data-services/1.5.2/installation-ecs/topics/cdppvc-installation-ecs-software-requirements.html

When we use browse-cluster.sh, it should launch SSB console to make it easier to access

Add "lite" labs from https://github.com/zBrainiac/edge2ai-workshop/blob/master/streaming_lite.adoc

After having an env for some time, there is no more /tmp/resources directory. It was there initially, as I ran it to prep for a demo.

What I was looking for here was the reset_to_lab.sh script

I assume it was deleted by normal OS cleansing routines.

My opinion is that /tmp/resources is not a good location anyway -- kind of hard to find.

Paywall credentials not being passed correctly to the cluster creation process for 7.1.9 stacks

Create lab exercises for Data Viz

When starting a page is listed http://tiny.cloudera.com/workshop-templates/

For this page to work, I believe you need to be on vpn.

However, when on VPN the url shortner does not work (consistently?).

Please link to the proper location directly

A bit of searching suggests https://github.infra.cloudera.com/araujo/workshop-templates

Ideas:

• Strip milliseconds
• Store timestamp as string and add conversion function

ECS instances should be managed with manage-instances.sh to stop and start using INSTANCE_IDS

CentOS 7 image used in workshop reached EoL - June 30, 2024

It was described that the demo boxes do not have many resources. If so, please instruct people to stop queries when they are not needed anymore, for instance after the section with 'run a simple query'.

Note this is about edge2ai-workshop/sql_stream_builder.adoc

The current permissions given to cdc_user are not sufficient to create a table in public schema.

The following works as a workaround, but may be too permissive. Needs to be reviewed and validated before going into the lab.
GRANT ALL PRIVILEGES ON DATABASE cdc_test TO cdc_user; \c cdc_test postgres GRANT ALL ON SCHEMA public TO cdc_user;

In case of failures during a launch, sometimes there are orphaned check-setup-status processes that keep running forever.

The login page asks for a password, which is often misleading for new registrants. We should change the label to "Registration code / Password"

The deployed Kafka service does not listen to localhost interface. For some tutorials, like Kafka Streams examples, the Java app is hard coded with the localhost parameter to reach out the Kafka broker.

I managed to configured to change a parameter, restart the service and listen to localhost (and all interface). Solution is described here: https://docs.cloudera.com/runtime/7.2.10/kafka-securing/topics/kafka-secure-multiple-listeners.html

Parameter value should be (and the one I used) is:

listeners=PLAINTEXT://:9092,SSL://:9093
advertised.listeners=PLAINTEXT://:9092,SSL://:9093

Requirements:

• Add Data Viz custom image in CDSW
• Allow unauthenticated access (do not require double authentication to access the Data Viz UI)
• Pre-create example dashboards / add support for reset-to-lab.sh scripts
• Update check-services.sh script

rpm -Uvh https://yum.postgresql.org/10/redhat/rhel-7-x86_64/pgdg-redhat-repo-latest.noarch.rpm

rpm -Uvh https://download.postgresql.org/pub/repos/yum/reporpms/EL-7-x86_64/pgdg-redhat-repo-latest.noarch.rpm

Installing k9s on ECS instance fails because yum provided golang package is lower than v1.21

This does not cause a launch to fail but it includes an unnecessary 5-minute wait for the SMON service to be ready.

Launches started to fail intermittently due to a timeout of the loading of Terraform plugins.
I'm not sure what changed. Maybe a Terraform update or my laptop upgrade to Sonoma.
Regardless, this should be more robust and retry for timeout-type failures.

When installing CEM 2.x it fails as CEM has now only jdk17 support:
https://docs.cloudera.com/cem/2.0.0/getting-started/topics/cem-system-requirements-efm.html

SSB-LB-UI needs libssl.so.1.1 provided by openssl11 package

Graph is outdated to an older version of CDSW.
The graph to extract the model_api_key needs to be refreshed to how it looks now.