Full Automated Install of RHEV-H?
We are trying to fully automated RHEV-H installation. We have found two issues so far that drive us nuts...
1) No matter what option used, install, reinstall or firstboot, the process stops and waits for operator to press enter? To confirm install? This happens every single time we boot via PXE. This makes no sense, how can we stop this behavior?
2) After automation takes place for RHEV-H deployment, the above issue not withstanding, we have to hit enter to acknowledge reboot? At least kickstart CFG had an option for automatic reboot, but not RHEV-H automation? No where in the Hypervisor Deployment Guide, or Installation Guide is this issue noted. Anyone figure out how to solve this one?
We have found a few other quirks, but these two just frankly, break the ability to deploy RHEV-H effectively, full automation? Not possible, yet.
Responses
Hi Schorschi,
What you are experiencing should not be happening at all, we deploy RHEV-H automatically from PXE all the time.
Can you post your ovirt.log and the PXE command line from a host that has gone through such a procedure?
In the RHEV-3.0-Node-02 entry try to use
APPEND rootflags=loop initrd=/images/RedHat/RHEV/3.0/initrd.img root=live:/rhevh-6.2-20120119.1.el6_2.iso rootfstype=auto ro liveimg nomodeset rootflags=ro crashkernel=512M-2G:64M,2G-:128M elevator=deadline processor.max_cstate=1 rd_NO_LUKS rd_NO_MD rd_NO_DM nocheck reinstall storage_init=/dev/sba BOOTIF=eth0 management_server=192.168.1.31 iscsi_name=iqn.2012-01.org.me.rhevnode02
BTW, storage_init=/dev/sba - shouldn't this be sda?
In my previous comment, I changed the line a bit more than just pointing to /dev/sba. There's a reinstall option in there.
I suspect you have some unclean disks in that host, and the install line you were using wasn't enough to cope with that.
Can you pls review my previous post and try that line exactly as it is (with /dev/sba-sda taken into account of course)
Does it work if you omit the iscsi_name param?
EDIT: Actually no, scratch that, leave iscsi_name in there, but also add local_boot
We have identified a potential bug, that can be worked around when local_boot is added in. Just want to verify you're hitting the same issue
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.
