No failover of MultiNic when both Nic on Node1 are down - Veritas Cluster Server

This is a discussion on No failover of MultiNic when both Nic on Node1 are down - Veritas Cluster Server ; I have a cluster environment with 2 nodes. Each node has 2 network connections. If one network connection fails, it should failover to the other connection. This works correct. When the two NIC's are down (I removed the network cables) ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: No failover of MultiNic when both Nic on Node1 are down

  1. No failover of MultiNic when both Nic on Node1 are down


    I have a cluster environment with 2 nodes. Each node has 2 network connections.
    If one network connection fails, it should failover to the other connection.
    This works correct.

    When the two NIC's are down (I removed the network cables) the cluster tries
    to failover from qfe0 to qfe4. That's normal. However, after the cluster
    is not able to bring qfe4 online I get the message "No more Devices configured.
    All devices are down. Returning OFFLINE".

    Instead of doing a failover to the other node, it tries to bring qfe0 online
    again on the same node.

    After about 10 minutes, the services do failover since the virtual IP address
    (a different resource) has been offline to long and it considers that resource
    to be Faulted. The virtual IP resource is depending on the MultiNic.

    I do want the services to failover to the other node so in the end, the result
    is fine, but I would think the cluster would do a failover because the MultiNic
    itself should be Faulted.

    Is this normal behaviour for the cluster?

  2. Re: No failover of MultiNic when both Nic on Node1 are down

    usually...MulticNICA will ping-pong between primary NIC and Standby NIC
    about 7 x before initiating a failover
    takes about max 2mins only
    possible to publish your MiultiNICA setting for me to take a look??
    "Petrie Hoek" wrote in message
    news:3d3fc20e$1@hronntp01....
    >
    > I have a cluster environment with 2 nodes. Each node has 2 network

    connections.
    > If one network connection fails, it should failover to the other

    connection.
    > This works correct.
    >
    > When the two NIC's are down (I removed the network cables) the cluster

    tries
    > to failover from qfe0 to qfe4. That's normal. However, after the cluster
    > is not able to bring qfe4 online I get the message "No more Devices

    configured.
    > All devices are down. Returning OFFLINE".
    >
    > Instead of doing a failover to the other node, it tries to bring qfe0

    online
    > again on the same node.
    >
    > After about 10 minutes, the services do failover since the virtual IP

    address
    > (a different resource) has been offline to long and it considers that

    resource
    > to be Faulted. The virtual IP resource is depending on the MultiNic.
    >
    > I do want the services to failover to the other node so in the end, the

    result
    > is fine, but I would think the cluster would do a failover because the

    MultiNic
    > itself should be Faulted.
    >
    > Is this normal behaviour for the cluster?




  3. Re: No failover of MultiNic when both Nic on Node1 are down


    In the types.cf file the settings for MultiNicA are:

    type MultiNICA (
    static int MonitorTimeout = 300
    static int OfflineMonitorInterval = 60
    static str ArgList[] = { Device, NetMask, ArpDelay, Options, RouteOptions,
    PingOptimize, MonitorOnly, IfconfigTwice, HandshakeInterval, NetworkHosts
    }
    NameRule = MultiNICA_ + group.Name
    static str Operations = None
    str Device{}
    str NetMask
    int ArpDelay = 1
    str Options
    str RouteOptions
    int PingOptimize = 1
    int IfconfigTwice
    int HandshakeInterval = 90
    str NetworkHosts[]
    )





    "Ricky Teo" wrote:
    >usually...MulticNICA will ping-pong between primary NIC and Standby NIC
    >about 7 x before initiating a failover
    >takes about max 2mins only
    >possible to publish your MiultiNICA setting for me to take a look??
    >"Petrie Hoek" wrote in message
    >news:3d3fc20e$1@hronntp01....
    >>
    >> I have a cluster environment with 2 nodes. Each node has 2 network

    >connections.
    >> If one network connection fails, it should failover to the other

    >connection.
    >> This works correct.
    >>
    >> When the two NIC's are down (I removed the network cables) the cluster

    >tries
    >> to failover from qfe0 to qfe4. That's normal. However, after the cluster
    >> is not able to bring qfe4 online I get the message "No more Devices

    >configured.
    >> All devices are down. Returning OFFLINE".
    >>
    >> Instead of doing a failover to the other node, it tries to bring qfe0

    >online
    >> again on the same node.
    >>
    >> After about 10 minutes, the services do failover since the virtual IP

    >address
    >> (a different resource) has been offline to long and it considers that

    >resource
    >> to be Faulted. The virtual IP resource is depending on the MultiNic.
    >>
    >> I do want the services to failover to the other node so in the end, the

    >result
    >> is fine, but I would think the cluster would do a failover because the

    >MultiNic
    >> itself should be Faulted.
    >>
    >> Is this normal behaviour for the cluster?

    >
    >



  4. Re: No failover of MultiNic when both Nic on Node1 are down


    You can control this behaviour of retrying by setting parameters

    Simer

    "Ricky Teo" wrote:
    >usually...MulticNICA will ping-pong between primary NIC and Standby NIC
    >about 7 x before initiating a failover
    >takes about max 2mins only
    >possible to publish your MiultiNICA setting for me to take a look??
    >"Petrie Hoek" wrote in message
    >news:3d3fc20e$1@hronntp01....
    >>
    >> I have a cluster environment with 2 nodes. Each node has 2 network

    >connections.
    >> If one network connection fails, it should failover to the other

    >connection.
    >> This works correct.
    >>
    >> When the two NIC's are down (I removed the network cables) the cluster

    >tries
    >> to failover from qfe0 to qfe4. That's normal. However, after the cluster
    >> is not able to bring qfe4 online I get the message "No more Devices

    >configured.
    >> All devices are down. Returning OFFLINE".
    >>
    >> Instead of doing a failover to the other node, it tries to bring qfe0

    >online
    >> again on the same node.
    >>
    >> After about 10 minutes, the services do failover since the virtual IP

    >address
    >> (a different resource) has been offline to long and it considers that

    >resource
    >> to be Faulted. The virtual IP resource is depending on the MultiNic.
    >>
    >> I do want the services to failover to the other node so in the end, the

    >result
    >> is fine, but I would think the cluster would do a failover because the

    >MultiNic
    >> itself should be Faulted.
    >>
    >> Is this normal behaviour for the cluster?

    >
    >



  5. Re: No failover of MultiNic when both Nic on Node1 are down

    what are settings ? it is necessary to change the arpdelay value to 1
    second, by right it is the default value. Shouldn't it probing the existing
    NIC and broadcast the IP will be less than 3 minutes ? I tested with the
    arpdelay 5 seconds, it does took about roughly 3 minutes.

    For your network hosts, you shouldn't use the IP that within the existing
    node. If your node A is using 192.x.x.A and node B 192.x.x.B, this 2 IP
    addresses shouldn't be in the NetworkHost, you should put in 192.x.x.x1 and
    192.x.x.x2, a different IP set !


    --
    Best regards,
    Effendy bin Yahaya

    "Simer" wrote in message news:3d6e707c$1@hronntp01....
    >
    > You can control this behaviour of retrying by setting parameters
    >
    > Simer
    >
    > "Ricky Teo" wrote:
    > >usually...MulticNICA will ping-pong between primary NIC and Standby NIC
    > >about 7 x before initiating a failover
    > >takes about max 2mins only
    > >possible to publish your MiultiNICA setting for me to take a look??
    > >"Petrie Hoek" wrote in message
    > >news:3d3fc20e$1@hronntp01....
    > >>
    > >> I have a cluster environment with 2 nodes. Each node has 2 network

    > >connections.
    > >> If one network connection fails, it should failover to the other

    > >connection.
    > >> This works correct.
    > >>
    > >> When the two NIC's are down (I removed the network cables) the cluster

    > >tries
    > >> to failover from qfe0 to qfe4. That's normal. However, after the

    cluster
    > >> is not able to bring qfe4 online I get the message "No more Devices

    > >configured.
    > >> All devices are down. Returning OFFLINE".
    > >>
    > >> Instead of doing a failover to the other node, it tries to bring qfe0

    > >online
    > >> again on the same node.
    > >>
    > >> After about 10 minutes, the services do failover since the virtual IP

    > >address
    > >> (a different resource) has been offline to long and it considers that

    > >resource
    > >> to be Faulted. The virtual IP resource is depending on the MultiNic.
    > >>
    > >> I do want the services to failover to the other node so in the end, the

    > >result
    > >> is fine, but I would think the cluster would do a failover because the

    > >MultiNic
    > >> itself should be Faulted.
    > >>
    > >> Is this normal behaviour for the cluster?

    > >
    > >

    >




+ Reply to Thread