Home > Flex-10, HP > HP Flex-10 DNS issue causing network outages

HP Flex-10 DNS issue causing network outages

March 4th, 2011

Updated with further information from HP:

HP has released an advisory explaining a serious issue with their blade Flex-10 Switches and DNS which can cause the Flex-10 switches to lose connectivity with each other in a Virtual Connect (VC) domain and cause network outages.

Apparently this is affecting a large number of customers although it seems to have taken some time to filter through the HP support system so if you are experiencing lost connectivity to your switches and network drops and have logged a call make sure the support engineer is aware of this advisory.

If you have DNS set up for any Flex-10 switches, there is a bug in the latest firmware, 3.15 (not sure if it affects previous firmware) which causes the Primary Flex-10 VC module to lose connectivity with the other VC modules.  This is apparently caused by the VC module initiating a reverse DNS lookup for the wrong IP address.

The workaround is to remove the DNS Server entries from the Flex-10 switch IP address settings so it doesn’t have any DNS servers to get the wrong IP information from.

It doesn’t matter if you are using DHCP or static IP addressing for your OA modules and/or Flex-10 switches as the manual settings override these so you can continue to use DHCP for the OA/iLO if are are currently doing so.

To remove DNS from the Flex-10 switches, in the OA for each chassis, go to the Enclosure Settings | Enclosure Bay IP Addressing, select the Interconnect Bays tab.

Enable Enclosure Bay IP Addressing (EBIPA) for each Flex-10 interconnect bay and manually enter the IP address, Subnet Mask, Gateway but leave the DNS Servers fields blank.

The Flex-10 switches will each need to be rebooted in turn for the setting to take effect.

I’ve updated my Flex-10 prerequisites post with this latest information.

The advisory says there will be a fix in a future firmware release.

Categories: Flex-10, HP Tags: , , ,
  1. June 8th, 2011 at 06:57 | #1

    Massive pain in the butt, this bug. It happened to me a few hours ago after I changed the OA module hostnames. Killed 3 blades worth of production VMs in one fell swoop.

  1. No trackbacks yet.
Comments are closed.