IFS and Salesforce

In plain wordsRecords from IFS are pushed into Salesforce, and changes made in Salesforce are carried back to IFS. Each direction is its own sync with its own configuration.

This page documents what the shipped catalogue sets up. For the commercial overview of the same pair — editions, licensing and how to buy it — Commercient publishes IFS to Salesforce integration on commercient.com ↗.

Phase 1 — IFS to Salesforce

What you getRecords from IFS become Commercient IFS Part Managed Custom Object, User, accounts, Commercient IFS Customer Managed Custom Object, Commercient IFS Address Managed Custom Object, Commercient IFS Item Warehouse Managed Custom Object, Commercient IFS Invoice Header Managed Custom Object, Commercient IFS Invoice Item Managed Custom Object, Commercient IFS Order Managed Custom Object, Commercient IFS Order Line Managed Custom Object in Salesforce. New records are created and existing ones updated.

Phase 2 — Salesforce back to IFS

What you getChanges made in Salesforce to accounts, Customer, Task, Activity, contacts, Order, Customer order, products, Part, Opportunity, Project, Contract, Sales contract, Service contract are carried back to IFS.

Setting up IFS and Salesforce from the CLI

One wizard stands the tenant up, and it is the same five steps for every pair: register, capture the server IP, connect the CRM, declare the ERP connector, then import and enable the sync processes. What changes between pairs is the CRM’s connection mode and the ERP’s connector type.

1. Register the tenant

Registration creates the account, emails a verification link, and — once the Data Lake is seeded — prints a once-only bootstrap API key valid for seven days. The three consent flags are the registrant’s own decision, and the server refuses start without all three.

# which instance types sign-up accepts - run this first and use IFS's code below
$ dlake register erps

# the three consent flags are the registrant’s decision - ask before passing them
$ dlake register start --generate-password \
$   --email <email> --company "<Company Name>" --phone <tel> \
$   --instance-type <code from dlake register erps> \
$   --consent-crm-backup --consent-erp-backup --consent-phone

# after the human clicks the verification link: watch for the seed and catch the key
$ dlake register status --watch

# the bootstrap key expires in 7 days - mint a durable one while it is still valid
$ dlake admin create_api_key --keyName "<tenant>-setup" --expirationDays 90

2. Connect Salesforce and declare IFS

Steps 3 to 5 of the wizard are admin-plane tools, they are safe to re-run, and the server enforces their order — attempt the CRM step before the server IP is captured and it answers 409 naming the step you skipped.

# the authoritative answer to where a registration got to - first call when resuming
$ dlake admin registration_state

# step 3: record the server IP (a prerequisite for the CRM step)
$ dlake admin registration_capture_server_ip --ipAddress <ip>

# step 4: the catalog lists every CRM with its authMode and the fields it needs
$ dlake admin registration_crm_catalog
$ dlake admin registration_crm_select --crmName Salesforce
# an OAuth CRM: the catalog’s callbackMode says which leg this one takes
$ dlake registration oauth --crm Salesforce
# this provider’s flow ends in a PIN, emailed by the complete call itself; the verb reads it hidden (or --pin-stdin)
$ dlake registration pin --crm Salesforce
# the connection is incomplete without this call, however healthy the rest looks
$ dlake registration action --crm Salesforce --action SalesforceSetup
$ dlake admin registration_crm_finalize

# step 5: the catalog lists the connector types and the fields each one needs
$ dlake admin registration_connector_catalog
$ dlake admin registration_connector_submit \
$   --connectorType <type from the catalog> --erpName <IFS code> --fields @fields.json
$ dlake admin registration_connector_provision
$ dlake admin registration_provisioning_status

3. Review and enable the Phase 1 processes

The shipped templates listed above are what crmpro_apply_template imports. Every one of them imports its process inactive, so an imported process does nothing until an operator activates it — read the process and its field mapping first, because an object with no mapped fields pushes nothing and records no error.

# the default-template catalog, and the mapping each template carries
$ dlake admin crmpro_templates --profile <tenant>

# what exists on the tenant now, and the master sync flag
$ dlake admin crmpro_list_processes --profile <tenant>
$ dlake admin crmpro_sync_status --profile <tenant>

# read a process and its field mapping BEFORE activating it
$ dlake admin crmpro_get_process --profile <tenant> --recordId <id>
$ dlake admin crmpro_field_mapping --profile <tenant> --recordId <id>

# turn the forward sync on
$ dlake admin crmpro_set_sync_enabled --profile <tenant> --enabled true

Processes run in a fixed order, and that order is a dependency order: a process that depends on another’s records sees them only once that process has run, so a parent flows on one run and its children on the next. The full order for this pair is on the Phase 1 page.

The wizard in full, including resuming a setup days later: Setting up an integration with dlake · Registering a new Data Lake · CRMPro (forward sync).

Source on GitHub

This page is generated from the skill pages in the dlake CLI repository. Those pages are the authority; if they disagree with anything here, they are right.

Related

Every ERP and CRM integration → · CRMPro (forward sync) · TxDownloaderPro (writeback sync) · AI agents & MCP · Salesforce integrations on commercient.com ↗