Last updated on APRIL 25, 2017
Applies to:Oracle Exalogic Elastic Cloud Software - Version 18.104.22.168.2 to 22.214.171.124.2
Oracle Virtual Server x86-64
In Exalogic Virtual Release 126.96.36.199.2 issue of unable to ping the guest vServers within the rack over EoIB address is observed. Issue happens intermittently across all the vServers.
Ping to EoIB address of the Guest vServers from the rack works without any issues. Only ping within the rack to EoIB of the guest vServer fails intermittently.
One other symptom observed when this issue happens is, when we look at the EMOC Job logs for stop and start of the guest vServers we do not see any tasks that indicate deleting and recreation of the VNICs. Usually when the vServers are started the VNICs are deleted and recreated by EMOC. In this case these Fabric tasks of deleting and recreating of VNICs does not happen.
All the network configurations and MTU settings are correct on the guest vServers. Upon digging further we observe that there are duplicate VNIC's for the guest vServers on the IB gateway switches. We deleted the duplicate GUID's and restarted the guest vServer from EMOC. When the guest vServer started again there were duplicate GUID's created on the IB Switches and EoIB network was not accessible within the rack.
Below are the steps which were followed for identifying the duplicate VNICS and deleting them.
Procedure for identifying and deleting the duplicate VNICs
- Get the GUID's of the guest vServers from the vm.cfg file. The GUID entry in the vm.cfg file will be as follows. In this example the rack is quarter rack, so we have 2 IB NM2-GW Switches. So we have two GUID's. In case of Full rack with 4 NM2-GW Switches we have 4 GUID's.
Sign In with your My Oracle Support account
Don't have a My Oracle Support account? Click to get started
Million Knowledge Articles and hundreds of Community platforms