pair of replicated ODBC ID sources (HA) - managing unreachable server scenario

[clandrai]

Hello,

This setup uses a pair of SQL DB with replication, each server is setup as separated ODBC ID source in ISE (2.3 patch 6). Policies are made so that if attributes can't get retrieved from ODBC1, ODBC2 source would be used instead (OR condition or doubled rules).

Problem: If one ODBC server fails, the rule will be impacted by a connection timeout, sometime leading to radius timeout with the NAD - and this for each authentication request.

Question: IS there an attribute similar to the MDM.isServerReachable that could be used to detect one server failure condition?

Alternatively, is there a plan to provide some HA, or which way is recommended at the server-side (unsure if MS SQL provides some way of HA)?

Many thanks in advance.

Christophe

[hslai]

ISE is not currently supporting an attribute to check for ODBC availability. Please connect with our PM team directly, due to No Comment on Roadmaps or Fixes in this forum.

[Armen V.]

Microsoft has implemented a connection string parameter to activate failover within the client called "MultiSubnetFailover". It seems however that there is no way to enter in that parameter within the ISE ODBC configuration.

https://docs.microsoft.com/en-us/sql/connect/odbc/dsn-connection-string-attribute?view=sql-server-ver15

We don't have control over connection string parameters to enable this within ISE, even assuming the driver underlying ISE is the microsoft-developed one.

[clandrai · Accepted solution]

Thanks for your reply, of course using a reachability attribute is just a possible way, whatever helps handling the loss and recovery of one server in a redundant setup is welcome - we will check what the SQL infra can offer, as well as solutions based on SQL clustering or load-balancers.

Thanks and regards,

Christophe