When the nodes install via pxeboot, they’ll retrieve the initrd and vmlinuz (kernel) via TFTP from the active controller. The initrd and squashfs.img must contain the same kernel and modules. The initrd loads up the kernel and modules, detects NICs, and retrieves the squashfs.img via HTTP. Once downloaded, it’s mounted and is the rootfs for the installer, with a chroot. The installer runs, Anaconda is launched and downloads the appropriate kickstart file. Anaconda installs all the required software, does various post-install steps, then reboots the node. At this point, the system boots from disk, running the installed kernel and software, and the installer images are not used.

 

 

From: Hu, Yong [mailto:yong.hu@intel.com]
Sent: Tuesday, September 18, 2018 12:12 AM
To: starlingx-discuss@lists.starlingx.io
Subject: [Starlingx-discuss] when and how is squashfs.img used?

 

As a part of pxe-network-installer-1.0-27.tis.x86_64.rpm, my understanding is that

Squashfs.img will be mounted by nodes (such as 2nd controller, compute nodes and storage nodes) during the process of PXE boot installation.

However, these questions are in my mind:

  1. Will be squashfs.img still being used *AFTER installation*?
  2. Where is squashfs.img mounted? I checked /etc/fstab, and there is no such an item.

 

Regards,

Yong