no valid license found on various AS Java NW 7.40 systems

[jonasmeyer1]

A lot of our AS JAVA NetWeaver 7.40 (approx. 40 systems) did an automatic shutdown. the defaultTrace said that there is no valid license for J2EE-Engine_ORA, although we know that the licenses were valid ones until 31/12/9999 (we still have hardcopies of the license files).

We checked the systems according to SAP note 1968213 and there were licenses present for J2EE-Engine_ORA and Maintenance_ORA on each system.

After starting the systems again, J2EE-Engine_ORA and Maintenance_ORA were marked as VALID.

We also already had a look at note 1898782.

To reproduce the issue, our Unix administrators migrated/moved the VM online/live to another KVM host and the hardware key didn't change. -> Date and time: 4.3.2017 appr. 10 am

In NWA we could immediately see that the license switched to invalid, but the hardware key was still the same.

We found this message in default trace:
Date 5.3.2017
09:56 am on Server1 and Server0 again the message: "J2EE-Engine_ORA" is NOT valid
10:26 am on Server1 and Server0 again the message: "J2EE-Engine_ORA" is NOT valid

After this message the SAP NW got stopped:
F [Thr 140332343318272] Wed Apr 5 10:26:06 2017
F [Thr 140332343318272] *** LOG => State changed from 3 (Running) to 4 (Waiting to stop).
F [Thr 140332343318272] *** LOG state real time: 88155.404 CPU time: 418.980 sys, 1004.070 usr
F [Thr 140332343318272] *** LOG total real time: 88231.674 CPU time: 432.630 sys, 1221.050 usr
F [Thr 140332343318272]
F [Thr 140332343318272] *** LOG => State changed from 4 (Waiting to stop) to 5 (Stopping).
F [Thr 140332343318272] *** LOG state real time: 0.000 CPU time: 0.000 sys, 0.000 usr
F [Thr 140332343318272] *** LOG total real time: 88231.675 CPU time: 432.630 sys, 1221.050 usr

With live migration of the VM we mean just a move from one KVM host to another KVM host, which happens in background (this means: we don't get informed by unix administrators, because there is/should be no
influence on the running SAP-system).

Our unix administrators described their procedure like this:

Live migration is started on the source VM Host using libvirt tools:
virsh migrate --live --verbose --undefinesource --persistent <VM_NAME> qemu+ssh://<Target_VM-Host>/system tcp://<Target_VM-Host>b

Additional information:
https://www.suse.com/documentation/sles11/book_kvm/data/sec_libvirt_admin_migrate.html

The systems of our landscape are running on KVM hosts of type HP DL580 Gen8 or Gen9 Hardware.
The OS of the KVM hosts is always SLES 11 SP4, while the actual VM is running on SLES 12 SP1:

SUSE Linux Enterprise Server 12 (x86_64)
VERSION = 12
PATCHLEVEL = 1
# This file is deprecated and will be removed in a future service pack or release.
# Please check /etc/os-release for details about this release.

NAME="SLES"
VERSION="12-SP1"
VERSION_ID="12.1"
PRETTY_NAME="SUSE Linux Enterprise Server 12 SP1"
ID="sles"
ANSI_COLOR="0;32"
CPE_NAME="cpe:/o:suse:sles:12:sp1"

Has anyone experienced a case similar to this?