RSA Auth agent not connecting to failed server with manual load balancing

[MaheshNune]

Hi,

We are using RSA Authentication Api 8.6 using C, with manual load balancing. Suppose we have 2 servers primary and replica, the api is trying to connect to primary and if it fails its switching over to replica. After sometime if replica is down but the primary came back up, The api is not trying to connect to primary again.

Its failing with error "No servers available"

Please let me know how to fix this?

Thanks,

Mahesh

#api authentication‌ #authentication manager sdk 8.6‌ #api 8.6.1 for c‌ #load_balancing‌ #sdopts.rec‌

[StevenSpicer]

Can you post a copy of the sdopts.rec from the affected agent host?  Please replace any IP addresses with "primary" "replica1" "replica2" etc so that your details aren't exposed here. Be careful not to remove any priorities.

[MaheshNune]

Hi Steven,

Please find the sdopts.rec file contents

USESERVER=primary,10
USESERVER=replica,1

Thanks,

Mahesh

[StevenSpicer]

Do you have a copy of the acestatus program to run on the agent host? If so, what does it show?

Does the agent log show any attempt to contact the primary after the replica goes down?  Have you turned the logging up to debug level?

You might need a more detailed investigation than we can provide in this forum.  I recommend you open a case with RSA Support so you can work closely with someone to resolve this.  https://community.rsa.com/docs/DOC-1294 

SIDE issue: A priority setting of 1 tells the agent not to use that instance for authentication unless all others (priorities 2-10) are unusable. Given that you only have two instances, is that what you intended?

[MaheshNune]

I had a case opened and waiting for some news on that. I don't have acestatus program, but a debug level set from my custom program we see the Manual load balancing is not trying to connect to Primary once it found replica is down.

In the aceclnt logs I see a message "  info AuthV8Session.cpp 1057 No servers reachable to send authentication requests"

I suspect the agent is somehow caching the server and its status and once its marked down its not retrying until we restart the agent. 

[StevenSpicer]

You're ahead of me, then.  Good. 🙂 You might call or send mail to support -- if you mail be sure to use the Case # in the title -- with your availability and contact info to arrange an online session.  https://community.rsa.com/docs/DOC-1294 

Good luck.