> I wonder if ever Dell answered this. Kind of. Have a look at the "Auto Dedicated NIC Feature in iDRAC7" http://en.community.dell.com/techcenter/b/techcenter/archive/2013/07/18/auto-dedicated-nic-feature-in-idrac7 http://en.community.dell.com/techcenter/extras/m/white_papers/20275980
Although this seems more to be meant for maintenance purposes rather than providing redundancy for a cluster fencing environment it should be usable for that kind of failover as well. Configured for "Shared" and "Auto Dedicated NIC" the iDRAC7 will failover to an active shared LOM (by default) 30 seconds after the link on the dedicated NIC goes down (tested, works). Careful with changing the default 30 second timeout if you have switches with standard 802.1d spanning tree timers (in a cluster environment you should anyway use 802.1w rapid spanning tree). Rather take care of the default timeout in your fencing logic. Stefan > On Jan 18, 2017, at 10:57 AM, lejeczek <[email protected]> wrote: > > > > On 17/01/17 19:00, Stefan M. Radman wrote: >> Hi Lejeczek >> >>> my iDrac6 currently runs off a dedicated NIC, and afaik it can be either >>> that or "shared". >> If your iDRAC6 has a dedicated NIC it's an iDRAC6 Enterprise. >> Available modes on the iDRAC6 Enterprise are: Dedicated, Shared, Shared with >> Failover LOM2, and Shared with Failover All LOMs. >> http://www.dell.com/support/article/us/en/19/SLN85572 >> >>> Thus, with "dedicated" not failover, am I right? >> With "dedicated" the switch your iDRAC6 is connected to becomes a single >> point of failure in your fencing logic. > > it must have begged the question - why Dell would not make "Dedicated with > Failover to..." > I wonder if ever Dell answered this. > >>> And one would not sacrifice "dedicated" for redundancy, would one? >> Yes, I would. What do you gain with "Dedicated"? >> Use one of the "Shared with Failover" modes and redundant network paths the >> same way you do with other cluster interfaces. >> >> Stefan >> >>> On Jan 17, 2017, at 6:32 PM, lejeczek <[email protected]> wrote: >>> >>> >>> >>> On 14/12/16 12:52, Stefan M. Radman wrote: >>>> Hi Lejeczek, >>>> >>>> The Red Hat Access website has two solution documents (content restricted >>>> to subscribers). >>>> One for rgmanager (RHEL5-6) >>>> How to configure fence_ipmilan with Red Hat Enterprise Linux High >>>> Availability Add On >>>> https://access.redhat.com/solutions/54843 >>>> and one for pacemaker (RHEL 6-7) >>>> How to configure fence_ipmilan with Red Hat Enterprise Linux High >>>> Availability Pacemaker Add On on RHEL 6, 7 >>>> https://access.redhat.com/solutions/2271811 >>>> >>>> One of my favorite open resources is Clusterlabs >>>> http://wiki.clusterlabs.org/wiki/Configure_Multiple_Fencing_Devices_Using_pcs >>>> >>>> The pacemaker solution has been working reliably here (RHEL6,HA >>>> AddOn,PE720) for more than two years now and actually triggered a few >>>> times, at least once during an actual hardware issue. >>>> >>>> Here is the stonith configuration we're using in a 2-node HA cluster >>>> running pacemaker on RHEL6: >>>> >>>> # pcs stonith create node1-pwr fence_ipmilan ipaddr=192.168.1.1 >>>> lanplus=true login=root passwd=calvin pcmk_host_list=node1 >>>> # pcs resource ban node1-pwr node1 >>>> # pcs stonith create node2-pwr fence_ipmilan ipaddr=192.168.1.2 >>>> lanplus=true login=root passwd=calvin pcmk_host_list=node2 >>>> # pcs resource ban node2-pwr node2 >>>> # pcs stonith show >>>> node2-pwr (stonith:fence_ipmilan): Started node1 >>>> node1-pwr (stonith:fence_ipmilan): Started node2 >>>> >>>> Make sure you have IPMI and redundancy (NIC failover) properly set up on >>>> your iDRAC. >>>> >>>> Stefan >>> oh, this is nice, many! thanks Stefan >>> my iDrac6 currently runs off a dedicated NIC, and afaik it can be either >>> that or "shared". >>> Thus, with "dedicated" not failover, am I right? >>> And one would not sacrifice "dedicated" for redundancy, would one? >>> regards >>> >>> >>>> On Dec 14, 2016, at 11:55 AM, lejeczek >>>> <[email protected]<mailto:[email protected]>> wrote: >>>> >>>> hi everybody >>>> >>>> just a quick question, probably very easy for you guys - >>>> iDracs as fencing device in HA configs(rhel's)? Would there >>>> be any docs/howtos? >>>> Or maybe there is better way to do that with PE servers? >>>> >>>> many thanks, >>>> L. >>>> >>>> _______________________________________________ >>>> Linux-PowerEdge mailing list >>>> [email protected]<mailto:[email protected]> >>>> https://lists.us.dell.com/mailman/listinfo/linux-poweredge >>>> > _______________________________________________ Linux-PowerEdge mailing list [email protected] https://lists.us.dell.com/mailman/listinfo/linux-poweredge
