![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
So here is the situation: We have an Active/Active 2000 SQL Cluster that has a single instance with SQL running on it (the other instance has been taken offline b.c. we are in the process of retiring the cluster). We also have an Active/Active/Active/Passive 2003 SQL Cluster that has two instances of SQL running on it (the other instance has not been implemented yet). The first cluster and the second cluster have simlar machine name but are in different domains. Here is an example of the setup: 1st Cluster node 1 - foo1.domain1.com node 2 - foo2.domain1.com Clustername - SQLCluster1 Instance name - SQLCluster1\foobar1 2nd Cluster node 1 - foo1.domain2.com node 2 - foo2.domain2.com node 3 - foo3.domain2.com node 4 - foo4.domain2.com Clustername - SQLCluster2 Instance name - SQLCluster2\foobar2 Instance name - SQLCluster2\foobar3 Notice that the nodes of the first cluster and two of the nodes of the second cluster have the same name EXCEPT they are in different domains. Okay so now that I've explained the situation, here is the problem; We are experiencing very odd performance issues on the SQLCluster2\foobar2. Where the CPU becomes throttled at about 40-50% and out websites become extremely slow. We are not seeing any blocking, and the connections are at a resonable number. I'm not exactly sure what the problem is, but one of my co-workers believes that because two of the machine in the 2nd cluster have the same name as the two machines in the first cluster we are experiencing issues with WINS/Netbios which is causing the CPU to become throttled. I do not believe this to be the issue. My experience is that when there are WINS/DNS/Netbios issues you will experience resolution or security issues but NOT performance issues. Anybody got any ideas/input? Thanks, Tom |
![]() |
| Thread Tools | |
| Display Modes | |
| |