I am challenged to migrate SPECTRUM 10.1.1 running on an Oracle T4-1 in a Solaris 10 LDOM to a VMware VM with RHEL 7 on an Intel platform. I am sure there will be no problems with the OS or the virtualization whatsoever. Nevertheless, I am being a little cautious in terms of hardware issues – while my current Solaris box was never a problem for CA and, with SPECTRUM 9.x, even preferred over a T2 or T3 driven machine (see https://support.ca.com/us/knowledge-base-articles.TEC593037.html), it seems that there are memory allocation problems (which, frankly, do not affect any of my other LDOMs, only the primary SpectroSERVER) that lead to CA refusing full support for performance problems we are facing with SPECTRUM 10.1.1. In particular, we have to restart our environment from time to time with weird side effects like corrupted SSDb (Stopping SS doesn’t work due to missing memory), thus the need to restore a backup which of course lacks the buffered events that were about to be written to DDMDb as soon as the stalled ArchMgr would work again… etc. pp.
We intend to buy Cisco UCM machines as a base for ESXi 6 – more precisely, UCM C240-M4 boxes (see https://www.cisco.com/c/en/us/products/servers-unified-computing/ucs-c240-m4-rack-server/index.html). We already use those for our CUCM environment. I assume we are less likely to be faced with an issue in the Intel world compared to Sparc, nevertheless I appreciate any input from the community in regards to possible problems somebody might have already faced with this particular hardware platform.