Our Setup: We have 2 sql servers in our primary datacenter and 1 sql server in our backup datacenter. We are using a different networks in each datacenter and the SQL listener has an IP for both networks.
All databases failed over to our backup datacenter node last night. (Still working on why they failed over in the first place)
Upon failover, a business application was unable to connect to its database. After looking in SSMS I noticed that DB's failed over to the backup datacenter node. I failed the DB's back to our primary node in our main datacenter. Within about 5 minutes, I was getting reports that users could access application again.
I decided to do some investigating. I did an nslookup and saw that both IPs were listed in DNS properly. The IPs were rotating if I ran NSLookup on the listener multiple times. So I then tried to ping the listener and noticed that the listener was still responding on the backup datacenter IP. I tried pinging multiple times and also flushed my local DNS cache. Each time the listener was trying to ping on the backup datacenter IP. I checked the cluster resource to see what the HostRecordTTL was set at. It was set at the default (1200 seconds). I adjusted it to 120 seconds. I waited a few minutes and tried again. Listener continued to use the wrong IP address.
It has been several hours since I initially started writing this post and the listener still seems to be trying to use the backup data center listener IP address. Our business application is working so I am assuming the listener is answering client calls on the correct IP. If that is the case, then why when I ping it do I still get the backup data center IP address?
All databases failed over to our backup datacenter node last night. (Still working on why they failed over in the first place)
Upon failover, a business application was unable to connect to its database. After looking in SSMS I noticed that DB's failed over to the backup datacenter node. I failed the DB's back to our primary node in our main datacenter. Within about 5 minutes, I was getting reports that users could access application again.
I decided to do some investigating. I did an nslookup and saw that both IPs were listed in DNS properly. The IPs were rotating if I ran NSLookup on the listener multiple times. So I then tried to ping the listener and noticed that the listener was still responding on the backup datacenter IP. I tried pinging multiple times and also flushed my local DNS cache. Each time the listener was trying to ping on the backup datacenter IP. I checked the cluster resource to see what the HostRecordTTL was set at. It was set at the default (1200 seconds). I adjusted it to 120 seconds. I waited a few minutes and tried again. Listener continued to use the wrong IP address.
It has been several hours since I initially started writing this post and the listener still seems to be trying to use the backup data center listener IP address. Our business application is working so I am assuming the listener is answering client calls on the correct IP. If that is the case, then why when I ping it do I still get the backup data center IP address?