first I try to give you some information. Our headquarter is located in Germany. All of our subsidiaries are connected to Germany via relatively slow VPN lines. Overall we have round about 20 DCs in different countires. Until now we have only 3 Palo Alto firewalls (Germany, USA, Canada) but in the future we plan to buy more.
Our setup until last week was the following. We installed the software User-ID agent on the DC in Germany. We connected all DCs worldwide to this agent. Unfortunately the agent retrieves all logs and produces a lot of traffic over the VPN lines. So we decided to change our setup.
We are planning to install the User-ID agent on all DCs worldwide and connect the firewall in Germany with all agents. With this solution the traffic should drop significantely. Furthermore the firewall in USA and Canada will be connected with the particular DC in that country and in Germany.
Or is it better to connect these firewalls also with all agents?
Is it possible to distribute the mappings from the firewall in Germany to the other firewalls without the need of connecting agents to the firewalls in USA and Canada?
Another question is if all agents should observe all available networks or only the local subnets?
I hope you can help me!
"We are planning to install the User-ID agent on all DCs worldwide and connect the firewall in Germany with all agents. With this solution the traffic should drop significantely."
Yes this will be fine.Also in the agents you may change 1second predefined read logs to 2(or 3) seconds.
Redistribute users is only available for agentless user-id.
to observer all available networks will be fine.
Currently we have 20 User-ID agents connected to our PA-2050. In the last days we had severe problems with our Palo Alto. The CPU was on full load and we were not able to login on CLI or the webinterface. The only solution was to restart the firewall. A few weeks ago we had the same problems but not as many as these days.
I'm not sure but something tells me that the integration of the User-ID agents is the cause of our problems. Is this possible?
The useridd process consumes a lot of memory. Is this normal?
The memory and cpu usage in the screen shot is not all that high. I would be more interested in the memory and cpu usage on the mgmtsrvr and devsrvr.
Run: show system resources follow
Use Shift + m
Per PanOS 4.1 and 5.0, 2050 supports 100 User ID agents.
Though it can support so many agents I would suggest the efficiency of the hardware with user id agents depends on the frequency of the data being retrieved from the agents to the firewall and how much data is being obtained. To some extent we can change the frequency values on the user id agents to retrieve user-ip mapping.
Hope this answers your question !
The mgmtsrvr server memory is border line but not extreme by any means. I would open a support case for the performance issue when it occurs again. This will need to be diagnosed live by a technician.
Click Accept as Solution to acknowledge that the answer to your question has been provided.
The button appears next to the replies on topics you’ve started. The member who gave the solution and all future visitors to this topic will appreciate it!
These simple actions take just seconds of your time, but go a long way in showing appreciation for community members and the Live Community as a whole!
The Live Community thanks you for your participation!