depot/modules/disk.nix, branch masterUnnamed repository; edit this file 'description' to name the repository.http://src.web.4kb.net/depot/atom/modules/disk.nix?h=master2026-08-31T18:43:56Z[nixos/disk] provide both qcow2 and raw images2026-08-31T18:43:56ZKleidi Bujarimail@4kb.net2026-08-31T18:43:56Zurn:sha1:2a06a6165be53ac383c3884a1a7383a3ecfb46c3[nixos/disk] experiment with autosizing zfs partitions in initrd2026-08-30T20:58:00ZKleidi Bujarimail@4kb.net2026-08-30T20:57:27Zurn:sha1:55dd84d67da8fb6e70aed7b921d05b66748546e1put depotImage along other generators and use naming conventions2026-06-07T03:52:09ZKleidi Bujarimail@4kb.net2026-06-07T03:52:09Zurn:sha1:30c548dbffe5ed489b0603c4d388a26a7aba4969write baseline config for new disk module2026-05-28T05:41:41ZKleidi Bujarimail@4kb.net2026-05-28T05:41:41Zurn:sha1:eacac6fa3eb4067b98b1d3ef84f9be2c6cacc660flatten dataset structure and persist more2025-07-06T00:40:45ZKleidi Bujarimail@4kb.net2025-07-06T00:40:45Zurn:sha1:08668f1e0e48a400cb577ff8228e0f3b2d20e56f
Modified options used in disk creation so that consumers of this
module will actually pass disko virtual machine tests.
The flat dataset structure is changed from the xnet disk module,
mainly since I never really made use of the nesting properly. The only
place nesting matters is really in the persist dataset, and that one
is managed by hand anyway.
Although the changes pass tests and seem fine, they have not been
tested on actual hardware yet since the dataset layout changes are
destructive to any running machines, and cannot be simply reverted,
since they alter the disk.
/var, /root, and extras are now persisted since they do not expect to
be wiped on reboot. I never really had problems with this but would
rather minimize messing with SystemD's expectations regarding /var especially.
automate hosts gen and bump nixpkgs2025-04-05T18:58:10ZKleidi Bujarimail@4kb.net2025-04-05T18:58:10Zurn:sha1:179bf31d5edc90c826dacf0b06b149469d8d14fa