Jazz Yao-Tsung Wang

The issue seems to be that our PXE phase uses legacy interface naming which makes the interfaces rather unpredictable, You're right. Based our experiment in 2009, it's the initrd-pxe which initialize all network devices (it's also when you see eth0 ... eth5). Your question reminds me that we tried to create eth0:1 for virtualization purpose (xen hypervisor) by modifing the initrd-pxe boot script. If you would like to know the booting steps of DRBL, here is a simplified animation: https://www.slideshare.net/jazzwang/08-06-24-drbl-introduction...

Hi Eric, Did you connect all NICs ( 6 NICs * 12 DRBL Client & 1 LAN NIC of DRBL Server ) to the same Network Switch/Hub? Let me try to answer each question in different threads. Q1. Why does the OS request DHCP on all interfaces? I'd like to understand: - Why does this happen after PXE has already configured the network via eth0? Quick answer: PXE only configure 1 NIC at a time. [PXE Phase] If you don't disable the PXE boot option of other 5 NICs, those 5 NICs will try to do DHCP and request PXE...

VOTE: clonezilla

VOTE: nsis

It's hard to tell why it disconnect. Maybe there are sshd "timeout" option in FreeNAS. Maybe there are some network limitation between your server and FreeNAS. An alternative way is use FreeNAS NFS instead of SSH. https://doc.freenas.org/9.3/freenas_sharing.html#unix-nfs-shares Jazz

Are you trying to backup /dev/sda1 to FreeNAS using sshfs ?? According to side FreeNas - error, there are error about 'sshd[5729]: _secure_path: /nonexistent/.login_conf'. and side CloneZilla live Ubuntu - error shows "Software caused connection abort". It looks like FreeNAS ssh disconnect ... so Clonezillla fail to store split backup files to FreeNAS. Jazz

Option 1 : use an external USB device for 256GB SSD Option 2 : save the backup files...

I found some discussion about NVMe grub compatibility issue http://forums.debian.net/viewtopic.php?f=17&t=128212...