LDV load performance: Bulk API and sharing deferral
Make tens of millions of records load fast: Bulk API 2.0 versus Bulk API batching and allocations, parallel versus serial mode and lock handling, sequencing roles, users, data and sharing rules, Public Read/Write and deferred sharing with the full recalculation, and trimming fields, automation and network overhead.
Compare and contrast various techniques for improving performance when migrating LDVs into Salesforce.
~8 min read · 18 practice questions ·
What the exam expects
- Choose Bulk API 2.0 or Bulk API for a large ingest, and explain batching, batch size and daily allocations.
- Pick parallel or serial mode, group children by parent, and handle lock failures and requeued batches.
- Sequence an initial load (roles, users, data, groups, then sharing rules one at a time) and use Public Read/Write or deferred sharing, ending with a full recalculation.
- Cut per-record work: only needed fields, triggers and flows bypassed for the load, clean data, session reuse, compression.
- Restart a failed Data Loader job without reloading what succeeded, and script repeatable loads with Data Loader batch mode or an ETL platform.
Key terms: Bulk API 2.0 (asynchronous API that batches and retries automatically); concurrency mode (parallel or serial processing of Bulk API batches); lock contention (batches competing for the same parent or sharing rows); defer sharing calculation (a Support-enabled permission to suspend sharing rule and group membership calculations); full recalculation (the sharing rule recalculation required after resuming).
Key facts
- Detail
- 15,000 batches per rolling 24 hours, shared by both APIs
- Detail
- 150,000,000 for each API; separate from Data Loader's 150,000,000-record CSV maximum
- Detail
- Records uploaded per day, visible at /limits
- Detail
- Start at the 10,000 maximum
- Detail
- Over 2,000 unprocessed requests from one org are delayed
- Detail
- Up to 10 reprocessing attempts; failed records not retried
- Detail
- Windows only; encrypted password and process-conf.xml
- Detail
- Not on by default; Customer Support enables it
- Detail
- Resume and run a full sharing rule recalculation
Sources for this section (5)
- Limits and Allocationsdeveloper.salesforce.com (opens in a new tab)
- Bulk API 2.0 vs Bulk APIdeveloper.salesforce.com (opens in a new tab)
- General Guidelines for Data Loadsdeveloper.salesforce.com (opens in a new tab)
- Command Line Introductiondeveloper.salesforce.com (opens in a new tab)
- Defer Sharing Calculationshelp.salesforce.com (opens in a new tab)
- Why can sorting a child file by parent ID avoid serial mode?
- When should a Bulk API job run in serial mode?
- Why are Data Loader's batch size settings irrelevant when Use Bulk API 2.0 is selected?
The questions are free to read. Answers with their sources and self-grading are included with the pack.
Keep reading with the Platform Data Architect pack
- Core concepts
- Exam traps
- Worked scenario
The pack includes:
- All 21 lessons of the Platform Data Architect guide, with reading progress
- Interactive decision walks and step-by-step sequence steppers
- Inline practice questions and one-click topic practice (18 questions for this lesson)
- An official source cited for every claim
- An AI tutor that answers from the guide
- Unlimited timed practice exams