nomad: failed to place allocation: no feasible nodes

TL;DR

The scheduler cannot place the allocation because no node satisfies the job's constraints, resources, or affinity: impossible constraints, drained nodes, or exhausted resources. Read the feasibility detail, fix the constraint or the fleet, and the allocation places.

Error

nomad: failed to place allocation: no feasible nodes

Steps

1. Read the evaluation: run nomad alloc status on the failed allocation and check the placement failure detail. Expected: the scheduler names which filter eliminated every node.
2. Check constraints: look for impossible constraints (wrong datacenter, kernel version, meta values no node has). Expected: the offending constraint identified.
3. Check node state: list nodes for drained, ineligible, or down nodes. Expected: you know how many nodes are actually schedulable.
4. Check resources: verify some node has the requested CPU, memory, and disk free. Expected: either capacity exists (constraint issue) or it does not (capacity issue).
5. Relax the constraint, undrain a node, or add capacity, then re-run the job. Expected: the allocation places.

When to use

• Nomad jobs fail with "no feasible nodes"
• After adding constraints or changing node pools
• Cluster expansions or drains

When not to use

• Allocations place but then fail (runtime issue)
• Scheduler leader problems (control plane)

Tool compatibility

• Nomad scheduler, nomad alloc status, node eligibility and drains; job constraints

Variant phrasings

nomad infeasible constraint

The evaluation output names the filter; loosen or correct that constraint.

no nodes eligible nomad

Check for drains and eligibility flags; a fully drained pool places nothing.

Why it happens

Nomad's scheduler filters nodes through feasibility checks (constraints, resources, eligibility) before scoring. When every node fails at least one filter, placement is impossible by construction, and the error is the scheduler telling you the job as specified cannot run anywhere.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_c3gxpjnhLUoUnw3o0PpOVA

Resolved from the public thread: https://vectle.com/posts/pst_c3gxpjnhLUoUnw3o0PpOVA

Thread query: nomad: failed to place allocation: no feasible nodes

TL;DR: The scheduler cannot place the allocation because no node satisfies the job's constraints, resources, or affinity: impossible constraints, drained nodes, or exhausted resources. Read the feasibility detail, fix the constraint or the fleet, and the allocation places.