Multiple UMP shared document_library?

[Emery Clark]

We are setting up multiple UMP servers behind a load balancer and came across this information about a Shared directory.

What is this shared directory used for? What happens if the primary UMP server goes down?

Notes below from: https://techdocs.broadcom.com/content/broadcom/techdocs/us/en/ca-enterprise-software/it-operations-management/ca-unified-infrastructure-management-probes/GA/configure-multiple-ump-servers.html

[Gene Howard]

the shared library is used by the Underlying Liferay engine to sync content between the two instances.
If the primary hub goes down the UMP will stop working as they can not authenticate users and the probes used for alarms and other things will not be available.
Currently, we do not have a support HA option for UMP.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Jesús]

Hello Gene

I have a question, Can I, use multiple UMP Server's (two UMP's) without load balancer configuration?. 
We have any problem with the portlets or to make a query on the portal with the users?

Release :

20.1  

The documentation that is being reviewed is the following:

https://techdocs.broadcom.com/content/broadcom/techdocs/us/en/ca-enterprise-software/it-operations-management/unified-infrastructure-management/20-1/installing/install-unified-management-portal-ump/configure-multiple-ump-servers.html

Thanks 
JLR

[Gene Howard]

Hi,
No this is not supported as changes made to one will not be properly synced to others.
Setting up multiple UMP without following the setup for this is not tested and so not supported.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Andrew Cooper]

JLR

We made a slight alteration to the architecture in the document to give a better active/active experience.

We have Linux hosts for the UMP. We also have a gluster cluster (3 nodes used for lots of stuff). so each UMP has its doc lib shared to the gluster :) thus we can stop any UMP without impacting availability from the LB and hence user perspective.

We originally setup up the lb as described in the doc, but I know that the LB admins experimented with other session persistence methods (source ip and such) as we found that doing tcp rather then app (http) balancing was faster and more reliable for the flash based interfaces (USM and such).  We do SSL offload at the LB as it keeps the UMP simple and we can troubleshoot the backend comms better.

Finally we are looking forward to the USM/flash replacement this month as that we will then be able to move to a more standard http(s) based LB.

Gene is right the primary hub is critical as this handles all the authentication for LDAP based users in the UMPs

Regards, Andrew

------------------------------
Knows a little about UIM/DXim, AE, Automic
------------------------------