- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Windows Authentication Agent still connecting to old primary appliance
Guys,
I need your help, I have replaced our replica appliance with a new unit, then promoted it to the new primary appliance (not changing naming) so now our old primary is now the replica and the new replica is now the new primary.
Now I need to remove the old primary which is now the replica to put in another new unit then make that the new replica and over time promote it back to the new Primary as its physical location is where all our servers are and the replicas physical location is at the Dr site.
When I switched the old primary off which is now the replica the windows authentication agents stopped authenticating as they are pointing to the old primary....
What is the correct way to now get all the Authentication agents to look at the new Primary to authenticate against? I cannot find any info in the manual AM 8.2
Thanks...
Hennie
- Tags:
- 8.2
- Agent
- Agents
- AM
- am 8.2
- Auth Agent
- Auth Manager
- Authentication Agent
- Authentication Manager
- authentication manager 8.2
- Community Thread
- Discussion
- Forum Thread
- primary replacement
- radius
- radius clients
- replica promotion
- replica replacement
- RSA SecurID
- RSA SecurID Access
- sdconf.rec
- SecurID
- Windows Authentication Agent
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
If these agents are Native SecurID Authentication agents (not RADIUS Clients), go to new Primary Security Console - Access - Authentication agents - Generate Config file (sdconf.rec) and download and unzip and
replace the current sdconf.rec on the agents that only go to the old primary replica.
IF they are RADIUS clients you have to manually configure them to use the primary or replica as the first RADIUS auth server and failover 2nd RADIUS server.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
If these agents are Native SecurID Authentication agents (not RADIUS Clients), go to new Primary Security Console - Access - Authentication agents - Generate Config file (sdconf.rec) and download and unzip and
replace the current sdconf.rec on the agents that only go to the old primary replica.
IF they are RADIUS clients you have to manually configure them to use the primary or replica as the first RADIUS auth server and failover 2nd RADIUS server.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
It depends:
you should not need to do much, but here is a full list of things to do/try
Verify packets on port 5500/udp are allowed from the agent to all RSA servers
and that it is not a network config issue preventing communication.
1) On primary...try a security console, access, authentication agents, auth manager contact list,
'automatic rebalance' on the current primary
this means the primary will only advertise the current state of instances to any agents
2) after rebalance, try distributing/replacing new sdconf.rec files to the agents
3) all windows agents also have a load balancing file called sdstatus.12, it
may contain stale server IP's. delete this file, the agent will be forced
to build a new one (automatically) that only has current server addresses
4) you may have a file called sdopts.rec on the agent machine. If there are
USESERVER statements in that file (if file doesn't exist no worries, it is optional)
then they may need to be changed
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks Jay your approach worked 100% , I just re created a new sdconf.rec file from the new Primary and copied the file to the location on Windows where the old file was and tested it 100%
Cheers man
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
To add to what Ed said, the automatic rebalance updates the contact list, which allows the primary to tell all agents about every replica. A contact list can be used to 'hide' replicas, maybe for development. I think they did this backwards but I'm just a Support guy. If you do the automatic rebalance before you generate the sdconf.rec file, when you generate all replicas will actually be stored in the sdconf.rec file, so agents with the sdconf.rec do not have to wait for the primary to tell them about replicas.
