HPE Synergy

IP address pool for servers iLO and Ether/FC interconnects

Occasional Visitor

IP address pool for servers iLO and Ether/FC interconnects

I try to desing HPE Synergy deployment. Now I'm thinking about IP-addresses for SY 480c servers iLO and interconnects modules.

I know that I can create ip-address pool at OneView settings page, but there is cpability for creating several pools. How to link one pool only for iLOs and second - only for ICMs? Is it possible?


Re: IP address pool for servers iLO and Ether/FC interconnects

In the standup of our Synergy frame, I do not see a way to do this. When you create the LIG and Enclosure groups, there is no way to specify more than one network. And when it applies to the Logical Enclosure, Oneview seems to assign the addresses from the pool at random. So I do not believe it is possible to have ILO seperate from the master VC interconnects. (possibly other interconnects since they would be in a seperate LIG)

Trusted Contributor

Re: IP address pool for servers iLO and Ether/FC interconnects

I had this same question on our Synergy installation this past week and we were able to accomplish this after initially setting three different pools of IP's and ending up with iLO IP's all over the place in a seemingly random grab from the available pools.  I like my iLO IP's to be assigned a bit more deterministically than that..  Here's what we did:

We started by removing the ranges we had entered in the Enclosure Group during the initial configuration and instead selectied the radio button for “Manage externally” in the Enclosure Group, then reapplied the Enclosure group to the Logical Enclosure (because it was now inconsistent) which generated a ton of errors from the iLO's.  I then removed the ranges we entered from Settings > Addresses and Identifiers section.  We already had a small DHCP scope on the iLO range and as I went into each Server Hardware int would show no IPV4 IP for the iLO, so I would Refresh the blade and it would grab one of the available DHCP addresses from my small scope in the iLO range (even though we did not specify use DHCP in the Enclosure Group but rather Manage externally).  I think this DHCP scope in the iLO range was critical to our success because until a blade grabbed an available IPv4 IP from my smal DHCP scope during the Refresh process, I could NOT open an iLO console session to that blade.  I then opened up the console from inside OneView and rebooted each host, hitting F9 to get into the iLO setup and changed the iLO name and IP to what we wanted them to be.  Then lather, rinse, repeat.  It was a tedious process but I got through all of our installed compute nodes in about a day,  and any newly installed nodes should be much easier to deal with going forward.

Although initially we were told that the ICM's and possibly other components like FLM also pulled IP's from the IP pool, we later got clarification that those internal devices all communicate with each other via internal IPV6 and do not rely on drawing IPv4 addresses from any pool(s) you might have created - this turned out to be true.  It took an entire day and every blade in the frames needed to be rebooted in order to enter F9 iLO setup to make these IP changes, but we did successfully change over from using IP pools to setting iLO IP's manually.

YMMV, and I would call support to validate your plan before attempting.