PowerHA for AIX

 View Only
Expand all | Collapse all

Power HA 7.1.3 SP3 under SAN boot

  • 1.  Power HA 7.1.3 SP3 under SAN boot

    Posted Thu September 24, 2015 06:01 AM

    Originally posted by: ThomasTse


    Hi all,

     

    I have the following settings:

    We have 2 * Power 8, Each Server are installed latest VIOs (version: V2.2.3.50) with Client LPAR (version 6.1 TL9 SP5), Ethernet heartbeat is set and OS disk, shared disk and repository are assigned via NPIV.

    During PowerHA UAT, ethernet failover is passed.

    However, fibre failover cannot be passed and the following is some scenario:

    1. unplugged fibre cable at server 1, server 1 will reboot and resource group will be failover to server 2 (successful scenario, but rarely occur).

    2. unplugged fibre cable at server 1, server 1 will not reboot and failover will not be occurred. once I manually restart server 1, server 2 will take over resource group. (most case)

    the above situation is applied when unplugging fibre cable at server 2.

    Can anyone suggest that what is happen? is it a bug? or is there any work around? Thanks.

     

    Thomas

     



  • 2.  Re: Power HA 7.1.3 SP3 under SAN boot

    Posted Mon September 28, 2015 04:12 PM

    Originally posted by: POWERHAguy


    Im not sure there is enough info. #2 is a bit confusing.

     

    In regards to #1, what it sounds like is happening is the rootvg system event is triggering (which was added in HA v7.1). But if you have redundant adapters, and multiple paths per, and are only pulling one, I'm not sure why rootvg loss is triggering. Make sure you're multipathing is working correctly.

     

    Also you can try by disabling the rootvg system event, to just log only and not reboot,  and see if you get different results. Check out section 9.4.1 of the (now discontinued, but I'm fond of it selfish reasons) HA7.1 redbook.

    https://www.redbooks.ibm.com/redbooks/pdfs/sg247845.pdf

     

    Overall something sounds a miss.

     

     



  • 3.  Re: Power HA 7.1.3 SP3 under SAN boot

    Posted Tue October 06, 2015 04:30 AM

    Originally posted by: ThomasTse


    Hi guy,

     

    Thanks for your information, we are doing UAT and unplug all fibre cables to see if it can failover or not.

    In regards to #1, I suppose it is normal behavior / actions taken to let server 2 to take over resources.

    #2 is abnormal behavior in my thought since it can't failover..

    Anyway, I already place a call to IBM for finding solution.

    Thomas