cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Dynatrace Managed - How to fix inconsistent service user and group identifiers among cluster nodes?

Radoslaw_Szulgo
Dynatrace Guru
Dynatrace Guru

How to change the system user on a specific cluster node?

When this node was deployed a while back, it was deployed with a completely different user than the other two nodes (3-node cluster). First, if I want to change the user, can I just run the reconfigure command with the --system-user-id  parameter? I want to change this node from userA to userB, so I am wondering if I can run this command with userB's UID and GID would that accomplish this? If that is the case, do I need to manually update any files (chown command) for the new user, or does the reconfigure command do this automatically? 

Senior Product Manager,
Dynatrace Managed expert
2 REPLIES 2

Radoslaw_Szulgo
Dynatrace Guru
Dynatrace Guru

1. Stop Dynatrace Managed on a node:

sudo /opt/dynatrace-managed/launcher/dynatrace.sh stop

2. Change uid:gid in the OS

usermod -u 2005 dynatraceuser
groupmod -g 3000 dynatraceuser

more help: https://www.cyberciti.biz/faq/linux-change-user-group-uid-gid-for-all-owned-files/


3. Execute reconfigure with the system-user-id param

/opt/dynatrace-managed/installer/reconfigure.sh --system-user-id <UID:GID>


4. Start dynatrace service:

sudo /opt/dynatrace-managed/launcher/dynatrace.sh start

 

Are there any risks with permissions for files breaking? Does this mean we also need to execute a recursive chown+chgrp to fix file permissions manually?

No. Reconfigure script will change permissions automatically.

Senior Product Manager,
Dynatrace Managed expert

AntonioSousa
DynaMight Guru
DynaMight Guru

@Radoslaw_Szulgo,

Thanks for this tip! I'll be having to do this next week in a client installation. I had checked in the documentation, and while it says all nodes should have the same uid:gid pair, it really doesn't say how to change that, if the Managed installation is unable to guarantee that in the first place. I would say that this tip of information should make it to the documentation.

I'll also join some notes that might be important for anyone reading this:

  • The main need for maintaining consistent uid:gid is mainly due to the backup process. Not sure if other shareable needs also require this...
  • There might be a case where we need to change all uid/gid in all cluster nodes (eg. where none of the uid/gid being used is available in all nodes). Is there a preferred way of changing them all?
Antonio Sousa

Featured Posts