Operator migration guide

Migrating PPPoE subscribers off Winbox and spreadsheets

Turn router accounts and spreadsheet rows into a reviewed service inventory. Move a small PPPoE group to RADIUS with one policy owner and a rehearsed way back.

Compare access models
Migration worksheet

Prove one group before expanding

policy owner
1
pilot group
1
review stages
6
PrepareInventory and plan mappingReview
PilotAuthentication and accountingVerify
RecoverPrevious policy and rateRehearse
Operator planning worksheet, not evidence of an automatic importer or a completed customer migration.
Page type
Guides
Primary search
migrate PPPoE subscribers from Winbox
Updated
2026-10-16

Winbox is useful for working on a MikroTik router. A spreadsheet is useful for reviewing a small customer list. The difficulty appears when the paid customer, PPP username and speed profile no longer agree across them. Moving to RADIUS should reduce that ambiguity before it changes how customers authenticate.

Start with the PPPoE billing scope. If you are also changing the access model, read PPPoE vs hotspot vs DHCP/IPoE first. Moving billing records and changing the subscriber's access protocol are separate changes with separate recovery requirements.

1. Inventory local PPP accounts without spreading passwords

Have the authorized router operator inspect /ppp secret, /ppp profile and active sessions. Build a sanitized worksheet; do not paste raw exports, passwords, shared secrets or backup contents into a ticket, sales form or shared spreadsheet. Keep recovery material in the operator's approved secure store with access restricted to the migration team.

Worksheet fieldWhy it matters
Customer ID and service IDDistinguishes a customer from their individual connections
NAS/site and PPP usernameLocates the exact access account; flag duplicates across routers
Account enabled state and hold reasonPreserves deliberate suspensions and disputed accounts
Profile, rate and per-user overridesCaptures what the subscriber actually receives
Address/pool assignment and CPE referenceReveals static-address exceptions and replacement hardware
Billing plan, balance timestamp and ownerEstablishes which finance record controls future decisions
Current session and last-seen timestampGives support a baseline for the first reconnect

Reconcile orphan usernames, duplicate customer rows, missing plans and accounts whose billing owner is unknown before selecting a pilot. Keep a versioned mapping and record who approved each exception. Do not infer payment status from an enabled router account.

2. Map profiles to commercial plans

A router profile is network configuration, not a complete billing plan. Map its intended upload/download rates, pool, session limits and special rules to a named service plan, then have finance review the price and billing period.

For example, a synthetic home-10 profile might map to HOME-10-MONTHLY. Two customers using that profile can still have different renewal dates or approved holds. Preserve those differences instead of converting every row to active on import.

Review per-user overrides and the default profile. RADIUS attributes can override defaults; settings not returned can still come from the default profile. Verify the result on the router. MikroTik defines rate directions from the router's perspective, so check both upload and download rather than copying a profile name. See PPP AAA profile behavior.

3. Prepare a separate RADIUS test identity

Agree the source of truth for customer records, balances, authorization and router configuration. Keep one writer for each. Register the correct NAS identity, confirm reachability and configure credential custody through the approved process. An endpoint that answers a health probe has not yet proven subscriber authentication or policy enforcement.

Before customer cutover, use an isolated test identity to prove:

  • Accepted and rejected authentication, with the intended profile/rate.
  • A session visibly authenticated by RADIUS, not a local user fallback.
  • Start, Interim and Stop accounting correlated to the same NAS/session; advancing timestamps and counters while test traffic flows.
  • Actual traffic and expected speed, checked separately from accounting.
  • The agreed primary/secondary outage behavior and a rehearsed return path.

The repository's PPPoE lab documents a narrow real PPP authentication/accounting proof. It does not establish public internet forwarding, Hotspot, payment-driven restoration or production readiness. Those remain separate acceptance checks.

4. Run in parallel without competing policy writers

Keep the existing customer population on its working path while a named pilot group moves to RADIUS. Compare billing and support records in read-only mode. Do not let two systems enforce overdue policies on the same service.

Local-account precedence is a cutover trap. RouterOS consults RADIUS when there is no matching local access record. A working login can therefore still be using the old local account. Have the router operator rehearse how the pilot identity leaves the matching local database and how it is restored during rollback. Do not assume disabling an entry implements timeout-only fallback. Verify the live session's authentication source. See the MikroTik RADIUS lookup rules.

Keep current balances and any manual changes reconciled throughout the pilot. Set the observation period around your actual billing events and staffing; there is no universal number of days that proves migration readiness.

5. Cut over one reviewed batch

  1. Agree the maintenance window, subscriber notice, approver and rollback owner.
  2. Capture the exact subscriber set, previous policies and rates, current accounting, and an accessible known-good router backup. Verify the recovery management path before touching authentication.
  3. Freeze competing policy changes for that set. Preserve intentional holds and suspensions; migrating is not permission to restore every account.
  4. Move only the approved identities using the rehearsed procedure. Verify authorization before refreshing sessions, then wait for each result.
  5. Confirm RADIUS authentication, fresh accounting, address, expected rate and customer traffic. Check an unaffected peer and reconcile finance records.
  6. Pause expansion on wrong identity, unexpected access, NAK, timeout, missing accounting or failed connectivity. An ACK or an active CRM badge alone is insufficient evidence.

Keep per-subscriber outcomes so support can distinguish migrated, unchanged, failed and rolled-back services. A timeout is ambiguous; inspect the NAS before retrying. Do not create a reconnect storm to clear a queue.

6. Rehearse rollback before the first customer cutover

This public-safe checklist summarizes the operator recovery approach. It is a manual, scoped procedure, not an ISPAgents automatic rollback capability.

  • Stop competing changes first. Identify the policy producer and account for queued or in-flight actions before applying a compensating change.
  • Restore one service at a time. Use the recorded old authentication path, authorization and intended rate. Reconnecting alone does not restore policy; restoring permission alone may not undo a speed limit.
  • Verify before proceeding. Check the exact identity and NAS, a fresh session, advancing accounting, actual customer traffic and expected rate. Preserve failures and unknown outcomes for the incident owner.
  • Use only a pre-approved emergency posture. Record scope, operator, expiry and recovery owner. A UI outage alone is not a reason to disconnect working subscribers. Do not enable a blanket permissive fallback.
  • Return deliberately. Recover the normal authentication/accounting path, prove a test identity, remove temporary bypasses in the reviewed order and reconcile billing. Preserve intentional holds and record residual work.

Use this rollback section as the public planning reference. The pilot team must maintain its own secure recovery packet and rehearse the exact router/firmware procedure; this guide cannot substitute for that proof.

Where ISPAgents fits today

Available foundation: basic subscriber records, search and device assignment.

Guided pilot: CRM/Customer 360, billing foundations, import preview/apply and coexistence records, service-policy previews and result recording, MikroTik RouterOS operations, and existing FreeRADIUS integration. Live mutation and recovery must be qualified for the selected tenant and access path. Real Splynx/UISP adapters are not a shipped integration promise.

ISPAgents Managed RADIUS remains guided pilot with tenant-specific onboarding, live smoke and launch signoff. Automatic PPP account import and end-to-end local PPP management are planned, not available capabilities. A parser or general import foundation does not establish credential migration, router write-back or a completed cutover.

For the next review, bring a sanitized profile-to-plan map and the pilot's acceptance checklist. See FreeRADIUS integration, WISP management scope, and the broader migration plan if an incumbent billing system is also part of your handover.

Frequently asked questions

Can ISPAgents automatically import all MikroTik PPP secrets?

Automatic /ppp secret import is planned, not a shipped promise. Begin with a sanitized inventory and operator-reviewed mapping. General import/coexistence foundations remain guided pilot; they do not guarantee PPP credentials or router configuration can be imported.

Can local PPP users and RADIUS run in parallel?

Yes, for deliberately separated subscriber groups. A matching local PPP user can take precedence over RADIUS. Prove RADIUS authentication for each migrated test identity; successful internet access alone does not prove the cutover.

Can the migration guarantee no downtime?

No. Reauthentication can interrupt service. Agree a maintenance window, test a small group, retain a recovery path and pause if identity, accounting, rate or connectivity checks fail.

Next step

Discuss the scope for your network.