Upgrades

[Adetweiler]

We took a deep dive in all settings in CMC in order to make sure we were prepared for the upgrade. We could not find of the settings in the release notes or the core documentation. It would be great if we could review the document before bothering your knowledgeable but very busy support team! We would be happy to be redirected to reading material.

[dylanwan]

We have the release notes for 5.1.x in https://docs.incorta.com/5.1/releases/

and 5.0.x in https://docs4.incorta.com/5.0/releases

May I know which settings are in question?

[KailaT]

Hey @Adetweiler, I wanted to make sue you saw Dylan's response. Which settings were you referring to. 

[Adetweiler]

Sorry for the delay I will have to check notification settings from the community mail.

This is what we couldn't find after reviewing docs, turns out some of this was old!

1. Tuning - Max versions per entity, backup frequency (as it relates to new dashboards/schemas)

2. Proxy read time out

3. clean up job refresh time.

4. Loader restart last columns (I guess that has always been there?)

These were questions resulting from the documentation that we couldnt self serve without support guidance (this was directed to support, so pls excuse questions)

1. In the documentation there is a note related to SSO that says configuration is required whether you are installing or upgrading but what would that be?
2. Overview of the new Lifecyle management of the schemas
3. A mention is made of the solution for long running dashboard queries. Can we review?
4. The docs mention that the loader service writing to full parquet instead of snapshot – can you confirm that is for Incorta Analytics as well as the sqli interface?
5. The new Automatic load order – do we need to remove our existing order
6. Reviewing the note below, is there a reason why would we disable this?

In the new mechanism, the engine forks only one thread for all columns required from a single physical schema object in the read request, whether a load job or a query. This feature is enabled by default for both types of read requests. However, you can disable it separately for each request type. You need to add a key to the engine.properties file that exists in the Loader Service directory or the Analytics Service directory, respectively. The path of this file is <installation_path>/IncortaNode/services/<service_directory>/incorta. The key that controls this feature is engine.read_parquet_columns_in_groups and you need to set its value to false to disable this feature.

In addition would be happy to have had a before/after output of CMC during upgrade, I had the team do screen shots. We were sure it was in the metadata somewhere but ran out of runway to look....

We have a list of upgrade bugs/issues that we have also shared, if you want review at some point as well! Have a good weekend - Amanda

[KailaT]

Thanks for sharing this @Adetweiler.

My team will be reviewing the documentation we have to ensure that the most accurate and current information is provided. We will also take a look at how to incorporate the questions you presented above into our documentation.

Please check your private messages so we can continue to chat about the bugs you have shared. 

In the meantime, I have followed up with Dylan to ensure your questions around settings following the recent release are adequately answered.

Thank you!