Collectors reporting "Registration failed: No Sumo credentials or access key provided"

Problem:

We have many collectors reporting:

ERROR com.sumologic.scala.collector.Collector - * ERROR: Registration failed: No Sumo credentials or access key provided

How do we get these collectors working again without having to reinstall the agent, and re-configuring them in the portal?

Cause:

This could happen for following reasons:

1.) When an existing collector loses it's credentials for whatever reason

2.) When the credentials gets expired in the backend for the existing collector because it was deleted or offline for 12 hours in case of ephemeral collectors, etc.

Solution:

The collector needs to be re-registered with the service with the steps given below:

1. You will need to update the user.properties file located at one of the following location:

   C:\Program Files\Sumo Logic Collector\config\user.properties (default for windows)

   /opt/SumoCollector/config/user.properties (default for linux)

  <collector-install-dir>/config/user.properties

• Make sure there is an entry for the "accesskey" in particular in addition to "accessid". By default the "accesskey" line will be removed from the user.properties file for security reasons unless a specific property is set with a line having  "skipAccessKeyRemoval=true"
• Check the remaining entries such access id (belongs to an active user), name field 

• If you have an existing collector configuration you want to mimic  you can download the json from Sumo UI as follows:

• In Sumo Logic select Manage Data > Collection > Collection. 
• Click the  icon to the right of the Collector .
  A dialog box opens to display the JSON configuration.

• Then you can add one of the following entries

sources=filepath or folderpath
syncSources=filepath or folderpath   # local json file monitored by the collector

•  If you are using the source.json file, note that there is a cutoffTimestamp setting for each source that indicates how far back in time the data should be ingested. If this is set to 0, then it will match all the (legacy or not)  files matched by the path expression. The cutoffTimestamp value is a Linux epoch timestamp that can be interpreted using epochconverter.com and this field can be set for every source so that data ingested does not result in duplicate data 

• If you want the collector to come up with the same name, delete the existing collector in the UI otherwise it will add a bunch of numbers to the existing name using the "<some_name>-epochtimestamp". This does not affect the data ingested by the collector refer this KB link  

2. Restart the collector service either via the Task Manager/services.msc in Windows or from the command line in Linux.

3. Verify the collector has successfully started up and is ingesting new data.