We have production issue with our sql 2K5 cluster. Our configuration is :
- 2 node IBM x3850 (active/pasive)
a) Node1 : IP1 (172.16.0.56) with mac address A, 4 nic teamed as public network - but only 1 is plugged into switch, 1 nic for private network
b) Node2 : IP2 (172.16.0.57)with mac address B, 4 nic teamed as public network- but only 1 is plugged into switch, 1 nic for private network
- OS : win 2008 enterprise R2
- DB : SQL 2005 SP3 enterprise
- Cluster Host IP (172.16.0.58)
- SQL Cluster IP (172.16.0.59)
SQL Cluster and Users segmented into different vlan. Server vlan is 172.16.0.x then User vlan is 172.16.4.x. Our sql cluster is functioning properly.
But when user accessed 172.16.0.59 from user's vlan, sometimes they got ping request time out and sql general network error. But not from servers vlan. We've checked the switch device, and showed that ARP table for sql cluster (172.16.0.59) is changing from
mac address A to mac address B. When the active node is node1, and the ARP table points node2's mac address, user got request time out.
Please help. right now the customer shutdown one of the nodes in order to get static ip-mac translation in arp table.
View Complete Post