When a network port (in MultiNICA configuration) is down, another port on local machine cannot takeover it, the float IP is up on the standby cluster node directly.

book

Article ID: 100005307

calendar_today

Updated On:

Description

Error Message

2010/11/02 23:15:03 VCS WARNING V-16-10001-6004 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Device e1000g1 FAILED
2010/11/02 23:15:03 VCS WARNING V-16-10001-6005 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Acquired a WRITE Lock
2010/11/02 23:15:03 VCS WARNING V-16-10001-6006 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Bringing down IP addresses
2010/11/02 23:15:04 VCS WARNING V-16-10001-6007 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Trying to online Device e1000g0
2010/11/02 23:15:05 VCS INFO V-16-10001-6008 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Sleeping 1 seconds
2010/11/02 23:15:06 VCS WARNING V-16-10001-6010 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Pinging 10.234.52.65 with Device e1000g0 configured: iteration 1
2010/11/02 23:15:11 VCS INFO V-16-10001-6008 (shc01aaaq01) MultiNICA:AAA-NIC:monitor:Sleeping 1 seconds
2010/11/02 23:15:13 VCS ERROR V-16-2-13027 (shc01aaaq01) Resource(AAA-NIC) - monitor procedure did not complete within the expected time.
2010/11/02 23:15:47 VCS ERROR V-16-2-13067 (shc01aaaq01) Agent is calling clean for resource(AAA-IP) because the resource became OFFLINE unexpectedly, on its own.
2010/11/02 23:15:48 VCS INFO V-16-2-13068 (shc01aaaq01) Resource(AAA-IP) - clean completed successfully.
2010/11/02 23:15:49 VCS INFO V-16-1-10307 Resource AAA-IP (Owner: unknown, Group: sh-Qchat-Oracle) is offline on shc01aaaq01 (Not initiated by VCS)
...

2010/11/02 23:16:30 VCS NOTICE V-16-1-10301 Initiating Online of Resource AAA-IP (Owner: unknown, Group: sh-Qchat-Oracle) on System shc01aaaq02
2010/11/02 23:16:34 VCS INFO V-16-1-10298 Resource AAA-IP (Owner: unknown, Group: sh-Qchat-Oracle) is online on shc01aaaq02 (VCS initiated)
 

Resolution

Check the network connection, we can see all of MultiNICA ports on two VCS nodes are connected to a Cisco Switch for public network application.

According to the NOTE as below:

It is recommended to set Spantree learning to "portfast" or disable the learning protocols options on host-connected ports (end-node devices, leaf-nodes) and on network switches that have this feature.

 

Try to enable "portfast" on the Switch, then do a test again, MultiNICA failover behavior is proper. 


Applies To

In main.cf


 IPMultiNIC AAA-IP (
  Address = "10.234.52.70"
  NetMask = "255.255.255.240"
  MultiNICResName = AAA-NIC
  )
 

 MultiNICA AAA-NIC (
  Device @shc01aaaq01 = { e1000g0 = "10.234.52.68",
    e1000g1 = "10.234.52.68" }
  Device @shc01aaaq02 = { e1000g0 = "10.234.52.72",
    e1000g2 = "10.234.52.72" }
  NetMask = "255.255.255.240"
  RetestInterval = 1
  )
 

Issue/Introduction

Two network ports are configured in MultiNICA, pull out the network link of the "running" port which has the float IP address; after a while, the float IP address is up on the VCS standby node directly but the secondary port on local node, although the secondary port on local node is not down.
Powered by Wolken Software