On Jan 24, 2013, at 12:18 PM, Máirín Duffy <duffy@xxxxxxxxxxxxxxxxx> wrote: > Thanks for the feedback guys! I made the changes you suggested except I > called the devices underneath "disks" instead of "drives" - since it'll > be a VM, I think they are called "virtual disks" so I thought using the > "disk" terminology might be more familiar. Does that make sense, or do > you still think "drive" is better? Coin toss. Doesn't really matter. "Drive" is a generic term for HDD, SSD, and even optical. Whereas "disk" indicates HDD, and "disc" indicates optical media. I'm going to guess that "disk" was casually usurped by VM designers prior to widespread use of SSD, so there was little meaningful distinction between drive and disk, hence "disk drive" as a term. Maybe drive sounds more like a physical object, and disk is more vague, so they picked disk. Really, the disk part is not at all emulated by a VM, but the drive (the interface, the command language, etc.) part is. So it's sorta amusing "virtual disk" and not "virtual drive" is common. > > Let me know what you think, you can shift-reload for the new version: > > http://linuxgrrl.com/fedora-ux/Projects/Anaconda/Usability/Tasks/RAID-Config.png Small coin toss on the (mirror copy) under the 2nd /home. I *think* it's clear that there are two instances of /home and that they are in a RAID 1 mirrored array without this. And with it, it seems like the 2nd one is somehow different, like it's the copy, compared to the first. When in reality they're identical at a block level. > > On Wed 23 Jan 2013 09:28:30 PM EST, Chris Murphy wrote: >> And honestly, I'd add two zeros onto everything. As long as > the virtual disk image isn't preallocated, and allowed to >> grow dynamically, it won't bust the volume its stored on. >> But such sizes are a more realistic test. > > You mean making all the sizes of everything bigger? The only thing is > by default, virt-manager doesn't do dynamic growing by default, so it's > one more thing you have to worry about when resetting the test for new > participants - and to delete the old VM's storage is another additional > click you have to remember - so I just want to make sure the test goes > as smoothly as possible and that there's no unintentional complications > that the usability tests will have to deal with. Yeah it's a small point then. Gnome Boxes does use dynamic disks by default. But I think it also does Express install by default too, so that has to be off every time, or you get a baked in kickstart install with zero control that starts immediately. But even if you just add one zero to each disk, you'd have to redo all of your size computations. Meh, probably not worth it. > Are there specific issues you think might crop up with the smaller > sizes that wouldn't crop up with a larger? The only one I know of is > that BTRFS has problems with really small disk sizes like this, but > we're not doing any BTRFS scenarios in the test. That's the main one I'm thinking of. Is this a LiveCD install or from DVD? If DVD, any concern that they go into Software Selection and pick a package set and options that need more space than 3.5 GB? Chris Murphy _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
