Is dracut-network installed?

You also might want to use the non-hostonly option for initramfs generation:

# dracut -N /tmp/initramfs.img

dracut-network is installed.

dracut -N /tmp/initramfs.img certainly ends up with everything but the kitchen sink, but no dracut iscsi
dracut-033-161.el7_0.173 with dracut modules:
bash
i18n
network
ifcfg
drm
plymouth
btrfs
crypt
dm
dmraid
kernel-modules
lvm
mdraid
resume
rootfs-block
terminfo
udev-rules
biosdevname
systemd
usrmount
base
fs-lib
shutdown

This is the header of the one built by anaconda iSCSI root, note the iscsi module.
dracut-033-161.el7_0.173 with dracut modules:
bash
i18n
network
ifcfg
drm
plymouth
kernel-modules
iscsi
resume
rootfs-block
terminfo
udev-rules
biosdevname
systemd
usrmount
base
fs-lib
shutdown

What is the output of dracut, if you add the "--debug" option?

Created attachment 1002506 [details]
Compressed Log of dracut debug output

dracut -N --debug initrd.img

There is more wrong with the migration that meets the eye. The VM loses it's iSCSI root disk overnight as well. Still responds to pings though.
Has done it each night so far :(

I enabled journald permanent logging but, of course, nothing was logged. No disk.

Are these rpms installed?
iscsi-initiator-utils
hostname

Sorry, that dracut did not tell you, that is was expecting to find:
- iscsistart
- iscsi-iname
- hostname

If one of these is missing, then it can't support iSCSI in the initramfs.

Current rawhide versions of dracut tell you what is missing.

(In reply to Darryl Bond from comment #6)
> There is more wrong with the migration that meets the eye. The VM loses it's
> iSCSI root disk overnight as well. Still responds to pings though.
> Has done it each night so far :(
> 
> I enabled journald permanent logging but, of course, nothing was logged. No
> disk.

dhcp lease expired? no dhcp daemon running?

> dhcp lease expired? no dhcp daemon running?
dhclient is running. The system lease time is shared for all iSCSI boot boxes, the ones that were built do not exhibit the behaviour. The fact that the VM can be pinged from outside indicates thet the network stack is still up


The interface doesn't have a timeout.
2: ibft0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:94:4f:f0 brd ff:ff:ff:ff:ff:ff
    inet 10.6.6.82/16 brd 10.6.255.255 scope global ibft0
       valid_lft forever preferred_lft forever



[root@localhost network-scripts]# systemctl status NetworkManager
NetworkManager.service - Network Manager
   Loaded: loaded (/usr/lib/systemd/system/NetworkManager.service; enabled)
   Active: active (running) since Wed 2015-04-15 07:05:27 AEST; 4min 50s ago
 Main PID: 723 (NetworkManager)
   CGroup: /system.slice/NetworkManager.service
           ├─ 723 /usr/sbin/NetworkManager --no-daemon
           └─1242 /sbin/dhclient -d -sf /usr/libexec/nm-dhcp-helper -pf /var/run/dhclient-ens192.pid -lf /var/lib/NetworkManager/dhclient-49abdca6-7015-4f8a-8335-bd59da6393f3-ens192.lease -cf /var/lib/NetworkManager/dhclient-ens192.conf ens192

Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info>   domain name 'gps.local'
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> Activation (ens192) Stage 5 of 5 (IPv4 Configure Commit) scheduled...
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> Activation (ens192) Stage 5 of 5 (IPv4 Commit) started...
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> (ens192): device state change: ip-config -> ip-check (reason 'none') [70 80 0]
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> Activation (ens192) Stage 5 of 5 (IPv4 Commit) complete.
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> (ens192): device state change: ip-check -> secondaries (reason 'none') [80 90 0]
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> (ens192): device state change: secondaries -> activated (reason 'none') [90 100 0]
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <info> Activation (ens192) successful, device activated.
Apr 15 07:05:31 localhost.localdomain NetworkManager[723]: <warn> Dispatcher script "/etc/NetworkManager/dispatcher.d/20-chrony" failed with timed out: Script '/etc/NetworkManager/dispatcher.d/20-chrony' timed out.
Apr 15 07:05:32 localhost.localdomain NetworkManager[723]: <info> startup complete

> Are these rpms installed?
> iscsi-initiator-utils
> hostname

It was missing iscsi-initiator-utils. That seems to have solved it. It generated a correct initrd with the Centos7.1 kernel update. That was without dracut -N, just the normal yum update. Some error messages like what is coming in rawhide will be good.

I will monitor to see if it continues to drop the disk overnight.

Dracut now shows info message about where is problem if it cannot boot with root on iSCSI target. This behaviour should be sufficient for user as notification to install correct packages and add right modules before creating initramdisk.img, which is able to use root filesystem on iSCSI target.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2016-2530.html