Diese Präsentation wurde erfolgreich gemeldet.
Die SlideShare-Präsentation wird heruntergeladen. ×

SFDX – Myth Buster, Svatopluk Sejkora

Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige

Hier ansehen

1 von 11 Anzeige

SFDX – Myth Buster, Svatopluk Sejkora

Herunterladen, um offline zu lesen

SFDX is an amazing tool, that allows for fast development, robust technology usage and great deployment performance.

During last years we have delivered multiple large projects and many smaller ones exclusively using SFDX.

We will have a look on how SFDX is currently working, what it is allowing to do and what are items that do NOT work as one would expect. We will go through real life scenarios that we have encountered during our projects and explain how we should be working with SFDX in a world of devs, admins, and business admins.

Topic 1 – Admins do work and get it to source control by one click
* Creating new field
* Deleting field / Renaming field
Topic 2 – Can profiles be managed by SFDX?
Topic 4 – Delta automatic deployment
* how to achieve them, what are the community tools to use
Topic 5 – Customer made changes to production, am I done?
Topic 6 – Metadata that doesn’t make sense and how to work with them
* we go through metadata that behaves differently and each user needs to understand how (translations, record types and many more)

I would love to have attendees leaving room with knowledge of what is possible and what are quick wins, what are gotchas and what to not try at all 🙂

Session will be guided with presentation, where all points will be explained and shown via examples within IDE on life scratch orgs.

SFDX is an amazing tool, that allows for fast development, robust technology usage and great deployment performance.

During last years we have delivered multiple large projects and many smaller ones exclusively using SFDX.

We will have a look on how SFDX is currently working, what it is allowing to do and what are items that do NOT work as one would expect. We will go through real life scenarios that we have encountered during our projects and explain how we should be working with SFDX in a world of devs, admins, and business admins.

Topic 1 – Admins do work and get it to source control by one click
* Creating new field
* Deleting field / Renaming field
Topic 2 – Can profiles be managed by SFDX?
Topic 4 – Delta automatic deployment
* how to achieve them, what are the community tools to use
Topic 5 – Customer made changes to production, am I done?
Topic 6 – Metadata that doesn’t make sense and how to work with them
* we go through metadata that behaves differently and each user needs to understand how (translations, record types and many more)

I would love to have attendees leaving room with knowledge of what is possible and what are quick wins, what are gotchas and what to not try at all 🙂

Session will be guided with presentation, where all points will be explained and shown via examples within IDE on life scratch orgs.

Anzeige
Anzeige

Weitere Verwandte Inhalte

Ähnlich wie SFDX – Myth Buster, Svatopluk Sejkora (20)

Weitere von CzechDreamin (20)

Anzeige

Aktuellste (20)

SFDX – Myth Buster, Svatopluk Sejkora

  1. 1. 1 Svatopluk Sejkora CzechDreamin SFDX - Myth Buster What you need to be careful about
  2. 2. 2 CzechDreamin
  3. 3. 3 Today’s Items 01 02 03 How to work as an admin on SFDX project Profiles <3 Delta package.xml for Deployment 04 05 06 Anyone changing metadata outside of agreed process SFDX metadata specifics Bonus
  4. 4. 4 Admin, Scratch Orgs and daily work It might sound intimidating, to start using console and commands and how the scratch org is even created. It is actually very simple, when you work closely within the team! • Utilize a plugin to make scratch org creation simple (envelope multiple commands) ► https://www.npmjs.com/package/ci-sfdx-plugin • Make sure all needed settings are applied and scratch org is ready to use once created ► You can find basic scratch org setup on the link below ► https://github.com/SvataSejkora/default-scratch-org-template • Renaming / Deleting items ► Search your repository for any instances, usual suspects: o Profiles o Permission Sets o Translations o Record Types
  5. 5. 5 Profiles When possible, do not use Profiles to anything besides: • Page layout assignment • Default app assignment • Desktop client access • Login Hours • Login IP Ranges • Session settings • Password policies • Delegated authentication • Organization-Wide Email Addresses • Default record type per object • Profile specific search layouts Reasons: • To get whole profile you always need to include everything when downloading metadata • SFDX pull can’t do it automatically • In large orgs profiles can have 15K+ lines easily What to do? • Go with Permission Sets and Permission Set Groups • This is going to be future anyways, but using source control is speeding the decision a lot
  6. 6. 6 Package.xml delta You never want to deploy all your metadata every time • It increases time to deploy • It increases potential of error There are plugins you can use, we are using sfdx-git-delta https://github.com/scolladon/sfdx-git-delta It works with GIT and will create package.xml from one commit to second. Usually, it is used to current HEAD from last successful commit. Does support generation from between of any 2 commits and can be used for redeployment of specific items. Either manual or automatic deployments using this tool are available and picking one will be based on your project and maturity.
  7. 7. 7 Changes outside of expected places No matter how robust process you have established at your project, there will always be situations when someone would make changes directly in UAT, Production or any other restricted environment. • Required minimum is to obtain information about the changes! ► Although there are options how to handle these situations without any knowledge, these are not covered in this session. SFDX has your back covered. You can, quite easily, obtain metadata from any connected org to your local repository: • sfdx force:source:retrieve -u connectedOrgName -m “MetadataType:MetadataName“ ► -m “Layout:Account-Account Layout“ ► -m “CustomMetadata:MetadataName.MetadataRecordName“ ► -m „CustomLabels“ • sfdx force:source:retrieve -u connectedOrgName -p src/main/default/objects/Lead/ ► Will download all items from within given folder
  8. 8. 8 SFDX metadata specifics It would be great to work with all metadata in the same way, but unfortunately it is not possible. • Custom Labels – always obtain all of them using *, otherwise you will need to merge a lot (fixed in the latest version) • Record Types – when pulling, include fields you have modified (fixed in the latest version) • Workflow Definition – should not be used, not supported anymore • Search settings – when you delete object you need to delete from here as well. Sometimes you will get a 0 in “resultPerPageCount” it can’t be zero for push. • Never include settings you are not using, they can be sources of many errors • Enabling tracking history for Record Type does not work. There is an error when pushing. Needs to be enabled manually if needed • Relate multiple contacts to activities cannot be reliable enabled in scratch orgs (as of May 2022) • AppMenus – do not include standard one, they cannot be reliable deployed. Also app switcher cannot be modified via API, so do not include it. Consider always using custom apps for you projects, and modifying app switcher manually. • Reports and dashboards – it is fine having them in source but you need to follow proper naming of your users for sharing (prod username + “.” + sandbox name. For initial scratch creation these names will be adapted to scratch org user name. This will not work for any subsequent pushes and will result in failure. Same goes for any user name reference in metadata (queues, flows)
  9. 9. 9 How to work with SF features and metadata You should be careful with your source control content. You need to have there anything you modified but items you are not using, do not need to be there. The reason is; they will be there when org or scratch is created as SF is providing templates, including standard items, for us. • There are items that you do not want to have in source ► Scratch org related items – they can exist only in production orgs, there is no advantage of having them (hub, Active Scratch Org) ► Items, you do not use within your implementation (ideas, solutions, knowledge, work…) ► Always try to minimize number of metadata you have (it influences time of scratch creation) • How to enable scratches with ► Knowledge ► Community licenses
  10. 10. 10 Any Questions?
  11. 11. 11 Thank you! #CD22

Hinweis der Redaktion

  • Show how the npm run sfci:init works and what all is then in Scratch: Data, assigned profile, assigned PSG
    Go through the items that are controlling all behaviour
    Project Scratch Def
    Ciconfig
    Package
    Create Custom field on account + translation and try to pull (it should not get all of the items)
    Rename the field for renaming, pull, show what is missing and explain why, and what to do
  • 1. Show after adding field to profile and downloading, that it is not here (needs to create first, download, then download profile - explain)
  • Show how the plugin works, how the deployments work.

    sfdx sgd:source:delta --to "HEAD" --from "HEAD^" --output .
    sfdx force:source:deploy -x package/package.xml -u targetOrg -l RunLocalTests sfdx force:mdapi:deploy -d destructiveChanges --ignorewarnings -u targetOrg
  • Show how to connect to an org -> sfdx force:auth:web:login -a CrestylProd -r https://test.salesforce.com
    Use created metadata (field, record Type, Translation)
    Show how to obtain them
    Talk a bit about formatting (Layout -, metadata without .mdt and so on)

×