Does it make sense to move all the products to XFS? An exception may be possible for cloud images [1] Server uses XFS. Cloud/Atomic have expressed a preference for it. I found that ext4 allocates inodes at mkfs time, and there's no way to get more later, and there's a real possibility with overlayfs that there can be inode exhaustion. [2] Yes, you can just use non-default mkfs options to create more inodes, but installer folks are typically reluctant to use non-default mkfs options. XFS and Btrfs have dynamically allocated inodes. Is there any advantage of LVM for Cloud/Atomic? What about Workstation? Dropping it for cloud/atomic may be an advantage, more compact, less wasted space, less overhead, one less layer. For Workstation it's probably a wash. Many users are confused by it, few benefit from it. But the documentation refers to it. --- Chris Murphy [1] Syslinux is preferred because it significantly reduces cloud image size. The problem is this extlinux + ext4 bug. While syslinux 6.04 supports XFS, and builds are in koji, it's not finalized yet upstream. So if this syslinux + ext4 bug gets fixed, then use use that combination. Otherwise, GRUB + XFS. Fedora 25 cloud images built with syslinux do not boot https://bugzilla.redhat.com/show_bug.cgi?id=1369934 [2] https://lists.fedoraproject.org/archives/list/cloud@xxxxxxxxxxxxxxxxxxxxxxx/message/47PC3YHBR7LTIDSDCZERROWRLEFOSVP7/ _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
