Hello,
I have a couple of questions about how the LDAP connection pooling work and I'm hoping that you folks might be able to shed a little light on it for me. We have 4 LDAP servers that are 'load balanced' via DNS using round-robin. Over the weekend one of the LDAP servers went offline. As soon as this happened, CAS stopped working and requests began to time out. At the time, I didn't know about the LDAP server being down and so, began to troubleshoot this issue as a problem with CAS. As a troubleshooting step, I bounced the service and it came up for about 10 minutes before the issues returned. Since CAS is pointed at a single DNS entry for the LDAP services, the nslookup returns 4 IPs. The order of the IPs rotates with each DNS query the DNS server receives. When CAS makes its connections, does it do an actual DNS lookup or does it only use the domain name:port thus relying on the network stack to handle DNS? How is CAS supposed to react in a situation where it cannot open a connection to an LDAP server? What is CAS supposed to do if a connection in the pool no longer works (ie. Starts timing out)? I also have a second context source that points to a completely different LDAP directory. This one is used to contain students and Alumni. Normally, if a user's not found in the first directory, it falls through to look here. However, during the failure CAS was timing out and not falling through. Is there a way to configure CAS so that it falls through to the second context source in the event that the first one times out? -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-dev
