Network Management
Remote collector not communicating after main poller upgrade due to version mismatch in SolarWinds Platform 2026.1
This article explains how to resolve a communication issue where a remote collector/additional polling engine stops communicating with the main poller after the main poller is upgraded to SolarWinds Platform 2026.1, but the remote collector remains on an earlier platform version (for example, 2025.2.1). It also documents an alternative resolution where the APE role is removed and the server is monitored as a standard node using WMI and/or the SolarWinds Agent.
First published date
Last published date
Overview
Observed issue:
After upgrading the SolarWinds main poller to SolarWinds Platform 2026.1, one or more remote collectors / APEs remain on an older version (for example, 2025.2.1). On the affected remote collector:
* SolarWinds services for the platform engine may be stopped or unstable.
* The engine appears down and not communicating in Settings > All Settings > Polling Engines.
* Attempts to run the 2026.1 scalability engine installer can fail with product catalog and Administration Service errors.
* The Administration Service log on the remote collector shows that it can read the database connection string but repeatedly reports that the local engine is not found in the SolarWinds Platform database.
Note: In some environments, direct internet access is not available on the remote collector, so the 2026.1 upgrade must be performed using the offline installer ISO.
Symptoms:
Typical symptoms reported:
* Remote collector and APE shows as Down or Unknown in Polling Engines.
* Devices previously polled by the remote collector show polling is failing or appear unmanaged / not updated.
* On the remote collector, platform services such as SolarWinds Information Service V3, Collector Service, and Job Engine v3 may be stopped or failing.
* The 2026.1 scalability engine installer reports:
“Could not download product information.”
“FailedToGetProductCatalogException: Could not download product information.”
* The remote collector's AdministrationService.log shows that it can build a valid SQL connection string to the SolarWinds Platform database, but repeatedly logs:
DB is not ready.
Local engine not found in DB.
This is valid state when the SolarWinds Platform is not configured yet.
Trying again in 30 seconds.
Environment:
* SolarWinds Platform main poller upgraded to 2026.1.
* Remote collector APE running earlier SolarWinds Platform (for example, 2025.2.1) at the time of the issue.
* Remote collector does not have direct internet access and is upgraded using a 2026.1 offline ISO.
* Remote collector was originally deployed as an Additional Polling Engine / Remote Collector and may also have the SolarWinds Agent installed.
Product section
Cause
The main contributing factors are:
1. Platform version mismatch
2. Incomplete or failed 2026.1 scalability engine upgrade on the remote collector
3. Engine registration not completed in the SolarWinds Platform database
Resolution
Resolution 1:
Option 1 is to rebuild the remote collector as a proper 2026.1 scalability engine (keep APE role)
Use this option only if the customer still requires this server to function as a remote collector / APE.
1. Verify database alignment
On the main poller, open C:\Program Files\SolarWinds\Orion\SWNetPerfMon.DB, and confirm the connection string points to the expected SolarWinds Platform database, for example:
Data Source=<SQL_SERVER_NAME>;
Initial Catalog=SolarWindsOrion;
Integrated Security=True;
2. On the remote collector, open C:\Program Files\SolarWinds\Orion\SWNetPerfMon.DB, and verify it points to the same SQL server and database as the main poller.
3. If the remote collector's file is incorrect or still shows an OLEDB/Access string, replace it with a copy from the main poller:
- Stop SolarWinds services and the SolarWinds Administration Service on the remote collector.
- Rename the local SWNetPerfMon.DB to keep a backup.
- Copy the SWNetPerfMon.DB from the main poller into C:\Program Files\SolarWinds\Orion\ on the remote collector.
- Start SolarWinds Administration Service and verify no “provider=” errors appear in the log.
Resolution 2:
Option 1: Prepare offline product catalog for the remote collector
- Mount the 2026.1 ISO on the remote collector.
- Ensure the following files and folders from the ISO are present on the remote collector under, C:\ProgramData\SolarWinds\Installers\Observability\RTM\
* catalog.json
* SubInstallers\... (all sub-installers)
- If necessary, copy these from the ISO to C:\ProgramData\SolarWinds\Installers\Observability\RTM\ so the installer does not need internet access to download catalog data.
3. Run the 2026.1 scalability engine installer on the remote collector
- On the remote collector, from the mounted ISO, right-click SolarWinds.Installer.exe > Run as administrator.
- Choose Scalability Engine and Additional Polling Engine.
- When prompted, point it to the existing main poller / SolarWinds Platform database.
- Allow the installer and Configuration Wizard to:
* Detect the existing 2025.2.1 platform on this host.
* Upgrade it to 2026.1.
* Register the engine in the SolarWinds Platform database as a valid polling engine.
4. Validate engine status
- On the remote collector, confirm in Programs and Features that SolarWinds Platform 2026.1 is installed.
- In the web console, go to Settings > All Settings > Polling Engines and verify:
- The remote collector appears in the list.
- Its Version matches the main poller (2026.1).
- Its status is Up and responding.
- Confirm that the AdministrationService.log on the remote collector no longer logs "Local engine not found in DB".
Option 2: Remove the APE Role and Monitor the Server as a Node
Use this option when:
* The remote collector role on the server is no longer required, or
* The 2026.1 scalability engine upgrade repeatedly fails or would require intrusive changes, and it is acceptable to perform monitoring directly from the main poller instead.
1. Uninstall Platform / APE Components from the Remote Collector
- On the remote collector server, Open Control Panel > Programs and Features.
- Uninstall all SolarWinds Platform components, such as:
* SolarWinds Platform
* Orion Scalability Engine
* Additional Polling Engine
* Other related SolarWinds Platform components
Important: Do not uninstall the SolarWinds Agent if it is already installed and will continue to be used for monitoring.
2. Keep the SolarWinds Agent and Treat the Server as a Standard Node
- Ensure the SolarWinds Agent service on the remote server is installed and running.
- From the Main Poller Web Console, Navigate to Settings > Manage Nodes.
- Add the server as a new node.
- Select the appropriate monitoring method:
* WMI polling, and/or
* Agent-based polling
- Provide the required credentials and test connectivity before completing the node addition.
3. After adding the server as a node via WMI and/or the SolarWinds Agent, validate that monitoring is functioning correctly. Ensure that the node appears in Manage Nodes with a healthy status and that system statistics such as CPU, memory, volumes, and interfaces are being polled properly. Additionally, confirm that the server is no longer listed as a polling engine under Settings > All Settings > Polling Engines, and verify that there are no further remote collector or engine registration errors in the logs on either the Main Poller or the remote server.
Additional notes and recommendations:
* Always verify that all scalability engines (APEs, remote collectors, additional websites) run exactly the same SolarWinds Platform version as the main poller after an upgrade.
* When upgrading remote engines:
- Run the Scalability Engine installer on the target engine itself, not on a jump server.
- For offline environments, always use the offline ISO and ensure the local catalog.json and sub-installers are available under C:\ProgramData\SolarWinds\Installers.
- If remote engine upgrade issues persist and the server does not need to function as an APE, it is often faster and safer to decommission the APE role and monitor the host as a standard node (Agent/WMI) instead of forcing a broken or partially registered engine to remain in the deployment.