What happens when a User-ID agent restarts?

[djr]

I have been using the agentless user-id but it seems to be overloading my firewalls so I am moving to a separate agent.  I am trying to decide whether I need one or two though and need to understand what happens when an agent restarts.

When it loses the agent, does the firewall drop all its user mappings?

When the agent comes back does it drip-feed new mappings or does it clear the existing mappings on the firewall from that source?

In other words, if my server were to reboot, would the firewalls carry on and pick up where they left off (accepting I would miss that any events while the agent were down) or would I lose all my user mapopings either when it went down or when it came back?

Most of my user mappings are from SYSLOG so are not re-generated.  Once they are gone, that's it.

[reaper · Accepted solution]

Hi

The existing mappings will remain active until they reach their idleTimeout/maxTimeout and will then be cleared if the agent is not available at that time, they will not be cleared just because the agent is unavailable.

When the agent returns it will reread the last 50k security logs on the ActiveDirectory and send updates to the firewall for any logins that occurred in those logs, nothing is cleared due to the agent returning to an active state. If any user mappings still exist on the firewall that require refresh, the firewall will ask the agent for information and it in turn can feed the firewall from it's stored list or perform a probe to determine a user. If most of your mapping is derived from syslog that may cause your user list to be depleted and eventually all mappings to timeout until a new syslog event is intercepted that can renew the mapping

hope this helps

Tom

[djr]

Great that sounds like it does what I want.  The only thing I didn't understand from your explanation was the "If any user mappings still exist on the firewall that require refresh,"  under what circumstances does the firewall ask for a refresh?

[reaper]

if the mapping times out(hard timeout, not idle) and the firewall has not received a refresh, but traffic is stil being received from the mapped ip, the firewall will request information from the agent regarding the user mapping

[djr]

That makes sense, but where are these two timeouts configured?  On my firewall and agent config I can only see a "user timeout".

[reaper]

there's also an idle and max timeout on the firewall, these are not configurable but rather depend where the mapping comes from (uidagent, captive portal, global protect, ...)

please take a look at > show user ip-user-mapping all

[djr]

Yes that's the command I usually use but I have to say I had not noticed the column headings!  For me the max and idle values are always the same so I had learned to ignore the last one.

Anyway that is great thank you for being so helpful,it has answered my question very thoroughly.