nfs crossmount ? - Aix

This is a discussion on nfs crossmount ? - Aix ; Folks, i have got 2 nodes hacmp: Node 1 owns a VG called data-vg in which resides only a LV (lv_data) mounted on /data/pub. /data/pub is defined in /usr/es/sbin/cluster/etc/exports and exported via nfs data-vg is a shared VG and is ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: nfs crossmount ?

  1. nfs crossmount ?

    Folks,

    i have got 2 nodes hacmp:

    Node 1 owns a VG called data-vg in which resides only a LV (lv_data)
    mounted on /data/pub.

    /data/pub is defined in /usr/es/sbin/cluster/etc/exports and exported
    via nfs

    data-vg is a shared VG and is "varyedoff" (standy) on node 2

    On node 2 there is an oracle instance that needs to access /data/pub
    (exported via NFS from node one)

    Is this possible ?

    Thanks 4 help!!


  2. Re: nfs crossmount ?

    hypnojazz@gmail.com schrieb:
    > Folks,
    >
    > i have got 2 nodes hacmp:
    >
    > Node 1 owns a VG called data-vg in which resides only a LV (lv_data)
    > mounted on /data/pub.
    >
    > /data/pub is defined in /usr/es/sbin/cluster/etc/exports and exported
    > via nfs
    >
    > data-vg is a shared VG and is "varyedoff" (standy) on node 2
    >
    > On node 2 there is an oracle instance that needs to access /data/pub
    > (exported via NFS from node one)
    >
    > Is this possible ?
    >
    > Thanks 4 help!!
    >

    Hi,

    yes this is possible and HACMP has a function for this. The important thing
    about this that you have to give HACMP control about the NFS crossmount. To
    accomplish this you have to do the following:

    - Define a new mount point (on both nodes) for the LV containing /data/pub (if
    data pub is not a separate filesystem you will have to accept that you will
    have a WARNING during cluster sync and verification, but it will work too).
    Let's assume you call the new mount point /data/pub_export

    - Change the HACMP resource group which manages the vg containing the lv in
    the field "Filesystems/Directories to NFS Mount" to the following entry:
    "/data/pub;/data/pub_export"

    These setting make HACMP to perform the NFS crossmount automatically for you
    on startup of the cluster (and remove the mounts during stop of the cluster
    manager). You should end up with the following mounts (numbers for sure are
    not yours):

    node 1 (active node)
    ===================

    Filesystem 512-blocks Free %Used Iused %Iused Mounted on
    /dev/publv 1920991232 315863448 84% 104650 1% /data/pub_export
    node1:/data/pub_export 1920991232 315863448 84% 104650 1% /data/pub


    node 2 (backup node)
    ===================
    Filesystem 512-blocks Free %Used Iused %Iused Mounted on
    node1:/data/pub_export 1920991232 315863448 84% 104650 1% /data/pub


    Regards,
    Uwe Auer


  3. Re: nfs crossmount ?

    Uwe,


    thank u very much!!

    I'll try your solution



    On Feb 15, 8:48*am, Uwe Auer wrote:
    > hypnoj...@gmail.com schrieb:
    >
    > > Folks,

    >
    > > *i have got 2 nodes hacmp:

    >
    > > Node 1 owns a VG called data-vg in which resides only a LV (lv_data)
    > > mounted on /data/pub.

    >
    > > /data/pub is defined in /usr/es/sbin/cluster/etc/exports and exported
    > > via nfs

    >
    > > data-vg is a shared VG and is "varyedoff" (standy) on node 2

    >
    > > On node 2 there is an oracle instance that needs to access /data/pub
    > > (exported via NFS from node one)

    >
    > > Is this possible ?

    >
    > > Thanks 4 help!!

    >
    > Hi,
    >
    > yes this is possible and HACMP has a function for this. The important thing
    > about this that you have to give HACMP control about the NFS crossmount. To
    > accomplish this you have to do the following:
    >
    > - Define a new mount point (on both nodes) for the LV containing /data/pub(if
    > data pub is not a separate filesystem you will have to accept that you will
    > have a WARNING during cluster sync and verification, but it will work too)..
    > Let's assume you call the new mount point /data/pub_export
    >
    > - Change the HACMP resource group which manages the vg containing the lv in
    > the field "Filesystems/Directories to NFS Mount" to the following entry:
    > "/data/pub;/data/pub_export"
    >
    > These setting make HACMP to perform the NFS crossmount automatically for you
    > on startup of the cluster (and remove the mounts during stop of the cluster
    > manager). You should end up with the following mounts (numbers for sure are
    > not yours):
    >
    > node 1 (active node)
    > ===================
    >
    > Filesystem * *512-blocks * * *Free %Used * *Iused %Iused Mounted on
    > /dev/publv * *1920991232 315863448 * 84% * 104650 * * 1% /data/pub_export
    > node1:/data/pub_export 1920991232 315863448 * 84% * 104650 * * 1% /data/pub
    >
    > * node 2 (backup node)
    > ===================
    > Filesystem * *512-blocks * * *Free %Used * *Iused %Iused Mounted on
    > node1:/data/pub_export 1920991232 315863448 * 84% * 104650 * * 1% /data/pub
    >
    > Regards,
    > Uwe Auer



  4. Re: nfs crossmount ?

    Hello everyone

    I had just upgraded from 5.3.0.3 to 5.3.0.6 and when I try to start both nodes I get the following:

    ERROR:
    NFS mount point /sapmnt/CRM is in VG vg_CRMIC in resource group crm_ic_rg.
    The NFS mount point must not be in a shared VG.
    FAIL

    ERROR:
    NFS mount point /sapmnt/CRM is in VG vg_CRMIC in resource group crm_ic_rg.
    The NFS mount point must not be in a shared VG.

    I used to get those as warnings in the previous version, but now I get them as errors and the nodes won't start. I had to reject the upgrade and go back to 5.3.0.3.

    Any ideas what could be going on?

    Thanks in advance!
    Lior

  5. Re: nfs crossmount ?

    Hello everyone

    I had just tried to upgrade a current, working HACMP 5.3.0.3 to 5.3.0.6 and when I tried to start both nodes, I got:

    ERROR:
    NFS mount point /interface is in VG vg_CRMIC in resource group crm_ic_rg.
    The NFS mount point must not be in a shared VG.
    FAIL

    ERROR:
    NFS mount point /interface is in VG vg_CRMIC in resource group crm_ic_rg.
    The NFS mount point must not be in a shared VG.

    and the cluster did not start. If I reject the upgrade (I applied the packages, not commit) and go back to the previous version everything starts working again and I only get warnings.
    Any ideas of what could be going on and how to fix it?

    Thanks in advance.
    Lior

+ Reply to Thread