lp:~smoser/curtin/trunk.install-deps

Branch merges

Propose for merging

No branches dependent on this one.

On hold for merging into lp:~curtin-dev/curtin/trunk

Server Team CI bot: Approve (continuous-integration) on 2015-12-09

curtin developers: Pending requested 2015-12-09

Diff: 233 lines (+206/-1)

3 files modified

curtin/commands/main.py (+7/-1)
doc/devel/README-vmtest.txt (+16/-0)
tools/install-curtin-deps (+183/-0)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

325. By Scott Moser on 2015-12-09

remove copied / bad code

324. By Scott Moser on 2015-12-09

fix pep8

323. By Scott Moser on 2015-12-09

fix usage

322. By Scott Moser on 2015-12-09

fix english

321. By Scott Moser on 2015-12-09

document patching images with curtin depends and provide tool

official vmtest should run with images as close to maas images as possible.
Thus, this should not be used for official tests.

However, developer use can patch the images that are used for boot, and
can dramatically decrease install time by installing curtin dependencies.

Example real world improvement even with a local proxy:
  # test before after
  PreciseTestBasic 66.22 37.79
  TrustyTestBasic 66.74 45.41
  VividTestAllindata 142.34 114.37
  VividTestBasic 76.89 50.03
  VividTestBonding 61.81 36.52
  VividTestLvm 73.47 48.32
  WilyTestBasic 78.65 53.53
  WilyTestBonding 64.48 38.70
  WilyTestLvm 87.24 51.62

320. By Ryan Harper on 2015-12-09

Use /proc/mounts to find missing mountpoints

lsblk (notably on trusty) does not find all mountpoints. Augment
block.get_mountpoints() lsblk data with mountpoints from proc/mounts.
Refactor reading proc/mounts into common function in block for reuse.

Fixes PreciseTestBasic (issues with btrfs on whole disk)
Fixes TrustyTestBasic (issues with dname and ptable)

319. By Scott Moser on 2015-12-08

tests: invoke nosetests with --nologcapture

When a install or boot fails, the vmtest harness would log.DEBUG
with the console log output. nose was capturing this logging
and replaying it on test failure.

This causes very long output to the screen of the user who ran
'make vmtest', making it hard for them to find out what happened.

Since we're now reasonably storing the console logs, its quite
sane to not spit those to stderr so that the output is easier to read.

318. By Scott Moser on 2015-12-07

vmtests: validate formatting/mounting of a disk without a partition table

Several use-cases involve formatting and entire disk with a filesystem.
Exercise this code by adding additional disks and tests to the test_basic
test. Utilize the btrfs filesystem type. Adding this test exposed several
issues with both fstab writing (it didn't know what to do with volumes of
type 'disk') and typo's in postinstall code which installed filesystem
utilities into the target (brtfs vs. btrfs). A couple other minor fixes
were included to ensure the testsuite

317. By Scott Moser on 2015-12-07

vmtest: do not expect sda-part2 to exist in /dev/disk/by-dname on trusty

This never worked, but we did not see failure before because of a test
case never properly executing.

316. By Scott Moser on 2015-12-07

dowait cleanups

make vmtest specify --dowait or --no-wait when running 'launch' and
'xkvm'. Additionally, have these programs be a bit smarter and default
to dowait=true unless both stdin and stdout are a terminal.
Previously, the default only considered stdin.

The issue that is fixed here is that if a developer ran:
  nosetests3 -vv tests/vmtests/
then xkvm would run qemu-system-x86_64 in the foreground.
by doing that, it would then not be able to handle the TERM signal
that it got from the timeout.