I am developing a long workflow, so I want to Pin a node near where I am working to avoid running the entire workflow multiple times.
Regardless of the number or location of Pinned nodes, the workflow always returns to the same sub-workflow near the start. The sub-workflow that execution always starts from contains an Information Extractor node. Is there anything about this node that requires it to always execute?
n8n always starts execution from a trigger node and runs through connected nodes up to the node you manually run. Pinning data only prevents re-running specific nodes, but doesn’t change the starting point or the execution flow. If you see the same early sub workflow running repeatedly, it’s because it’s on the path from your trigger to the node you’re testing. Pinning just reuses output but doesn’t stop the workflow from starting there.
How to avoid re-running the starting steps
Add a Manual Trigger right before the part you’re working on.
Connect it directly to your focus area, bypassing the early sub-workflow.
During testing, start execution from this new trigger, so the big initial steps are skipped.
Pin the nodes you want to keep stable to prevent re-executing slow or costly steps.
This is exactly the behaviour I am expecting, but I am not seeing it.
I have multiple nodes with pinned data, including the long-running node. Each time I execute the workflow or execute a single node, we reach the long-running node and have to wait minutes for it to complete. The pinned data is ignored.
What can cause pinned data to be ignored? Is it possible to pin output from a sub-workflow?
@ajf Pinned data only prevents re‑running nodes during manual, editor‑based tests. It is a development feature and is ignored during actual workflow executions triggered by webhooks, schedules, or parent workflows. Execution always starts from the trigger and proceeds forward, so pinning only reuses data at a specific node and doesn’t skip earlier steps or sub‑workflows called during production runs. To test with pinned data, develop using a manual trigger or run sub‑workflows directly in the editor with nodes pinned beforehand. This ensures pinned data is used, as pinning isn’t active in production-style executions. Let me know if this helps
No @Isikyus_BN of course we can run a single node in the editor, but when we are building a complex flow such as ‘Webhook’ → ‘Email Auth’ → ‘Save On Sheets’ , we need to require pinning the data inside the webhook when we first tested it and so we can take reference on how that data would process in the flow.
Like so often happens, posting on a support forum helped me focus and see my oversights.