The userid process suddenly stops working and stops running.

User-id mapping entries are not being synchronized to secondary node and "show services user-identification identity-management status" command shows "error: the user-ad-authentication subsystem is not running"

If userid mapping entries are not being synchronized between nodes and userid process suddenly stops working with "error: the user-ad-authentication subsystem is not running" error check the following:

• show system processes extensive | match userid <---- Get userid process PID, memory utilization and CPU usage
• show system storage <---from CLI to see storage usage on the device
• If you notice high storage utilization from the above command then from shell run "find / -size +50M | xargs ls -lhS" to look for large files on the system
• If you notice any file under /var/db/userid taking too much storage then this may be the reason for the failure and for the process not working as usual, usually these files are /var/db/userid/auth.sdb.repl.sdb and /var/db/userid/auth_v2.sdb.repl.sdb

Once the file taking too much storage is found you can delete it from shell using "rm <path to filename>", example: "rm /var/db/userid/auth_v2.sdb.repl.sdb".

Once the file is deleted check the storage and see if it is now normalized, if so then the next step would be to kill the user id process, for this we will need the PID:

1. Get PID of userid process from shell: "top -H | grep userid"
2. Forcefully kill the process from shell: kill -15 <PID>

Once this is performed check for the status of identity management.

This issue is tracked under PR1701990, check https://prsearch.juniper.net/problemreport/PR1701990.