~wesley-wiedenmeier/curtin/install-kernel : changes

~wesley-wiedenmeier/curtin/install-kernel » Changes from revision 346

From Revision 346 to 327

Rev	Summary	Authors	Date
346	Merge from lp:~smoser/curtin/trunk.install-kernel-fix Merge from lp:~smoser/curtin/trunk.install-kernel-fix	Wesley Wiedenmeier	8 years ago
345	Curtin sometimes cannot detect the correct kernel package to... Curtin sometimes cannot detect the correct kernel package to install based on build in kernel mappings. If this happens, instead of no longer attempting kernel installation, try to determine the correct kernel to install based on the kernel version referenced in the linux-image-{flavor} meta package. This allows curtin to do the right thing when installing using a ephemeral image that does not contain a kernel in the event that no kernel is specified in curtin's configuration. Previously curtin would attempt to determine the correct kernel to use based on the os.uname() of the system running during the installation, but this fails when the host system is using a kernel version not specified in curtin's build in kernel mappings, such as when installing a xenial ephemeral using wily. Curtin will still attempt to use the built in mappings, and only fall back to this detection if that fails.	Wesley Wiedenmeier	8 years ago
344	mkfs: fix for lack of uuid on precise or trusty mkfs: fix for lack of uuid on precise or trusty the last merge for mkfs broke vmtest. Precise's version of btrfs lacks both uuid and force flag. Trusty's lacks uuid.	Scott Moser	8 years ago
343	Improve mkfs command and move mkfs logic to block.mkfs Improve mkfs command and move mkfs logic to block.mkfs The 'mkfs' command now operates as a useful utility for users who want a consistent interface to 'mkfs'. The mkfs logic is improved and moved to the block.mkfs module to make it more standalone.	Wesley Wiedenmeier	8 years ago
342	vmtests: basic_network.yaml remove static gateway to avoid r... vmtests: basic_network.yaml remove static gateway to avoid race with dhcp The basic network configuration includes a static gateway which will race with DHCP; last one to write wins and sometimes means the network validation won't find the static gateway and fail. Remove the static gateway in the presence of a network configured for DHCP.	Ryan Harper	8 years ago
341	vmtest sync: fix pruning of dirs vmtest sync: fix pruning of dirs our pruning code is somewhat simplistic at this point. We really need to have a better mirror, but for now this fixes an issue. At some point on our c-i nodes, a mirror included 'di' data, which creates files in '<mirror>/<release>/<arch>/di'. That directory was getting counted as a 'image dir'. This will count image dirs as those that are of format YYYYMMDD[.N]	Scott Moser	8 years ago
340	tests/vmtests: fix invalid variable and run pylint on tests/... tests/vmtests: fix invalid variable and run pylint on tests/vmtests This fixes a bad variable in tests/vmtests anad also starts to run pylint with python3 on tests/vmtests. pylint would have caught my bad variable. Also fix the other complaints from pylint on that code so that it will successfully run.	Scott Moser	8 years ago
339	maas2roottar: use --numeric-owner when creating root.tar.gz maas2roottar: use --numeric-owner when creating root.tar.gz We've just been lucky to have not hit issues with this. Since the image may have different user and group numbers than the host system, we need to use --numeric-owner with tar.	Scott Moser	8 years ago
338	Further cleanups on syncing images Further cleanups on syncing images This clears up some of the unexpected behavior of syncing images. Much of the confusion was because sstream-mirror keeps internal information in .data/ that sstream-query does not read. The result is that: sstream-mirror <remote> <local> <filters> and then a sstream-query <local> <other_filters> would show results as if <other_filters> had been mirrored even though the files were not present. We clean that up. Now, get_images will ensure that the data you were after is present. Also, we no longer attempt mirror on every setUpClass. That could result in one Precise test running with one image and the next Precise test running with another image. This will only run a mirror once for a given set of filters during a nosetests run.	Scott Moser	8 years ago
337	sync-images: auto-convert and prune images sync-images: auto-convert and prune images make sync-images will now ensure we invoke convert_image() as needed on each one and we now have a default of 1 image to keep around. This number is controlled by an environment variable, IMAGES_TO_KEEP.	Ryan Harper	8 years ago
336	mdadm: fix issues exposed by use via block_meta mdadm: fix issues exposed by use via block_meta Using mdadm via block_meta exposed a few issues. - block_meta calls block.md_check, this should be mdadm.md_check - mdadm.sysfs_read_attr had thinko in check if it should attempt to read an attribute and avoiding reading files - mdadm.sysfs_read_attr should use util.load_file - ensure all entrypoints to mdadm code check that device names are device paths.	Scott Moser	8 years ago
335	add running of pylint for python27 and python34 add running of pylint for python27 and python34 This adds the running of pylint in '--errors-only' mode for both python27 and python34. pylint is very verbose, but in --errors-only mode, it will catch bad many errors that would not be caught by flake8 without unnecessary noise. It finds a couple legitimate issues, and I've included fixes here for those. There is one work around for an issue i opened at https://github.com/PyCQA/pylint/issues/769	Scott Moser	8 years ago
334	initial fix for reading skew_data initial fix for reading skew_data I picked these from cloud-init revno 1147 that made similar changes. One thing I don't yet have here is the retries that went in there. Essentially you have to retry to put the skew in place. In cases where MAAS is running curtin, cloud-init will have already done that for us and we'll just read the offset from skew_data_file. So this should be sufficient for MAAS, which is really the only place that OAuth is currently used. There is actually one other set of changes that i'd like to get. Currently we set the oauth skew, and then use it, but if something (systemd-timed) updated the clock during these things happening then our skew could be come wrong. This is described more in bug 1531233.	Scott Moser	8 years ago
333	support bcache installation on precise support bcache installation on precise We can now do bcache installations on precise with the hwe-t kernel. - list bcache-tools package as a dependency on precise Bug 1449099 covered SRU of bcache-tools to precise. bcache-tools is now available in precise-updates. - Bug 1513176 added bcache to the initramfs provided for maas images. I've added a simple bcache installation test here and enabled it for trusty and for xenial. It is not enabled for precise because we do not have ability to test hwe kernels yet. Also, fix a bug in assumption that the id of a bcache device was the kernel name for that. The bug was exposed in this bcache_basic.yaml config.	Scott Moser	8 years ago
332	fix bug in install_grub to partition when storage_config pro... fix bug in install_grub to partition when storage_config provided. when installing with a storage config and grub is to be installed on a partition (as it is in powerpc) we were not able to look up the device that was to be installed on. This puts a temporary bit of code in place that finds that device via imports from commands.block_meta. The result is that storage_config now works for lvm on power8 systems.	Scott Moser	8 years ago
331	url_helper: raise import error on lack of oauth only when oa... url_helper: raise import error on lack of oauth only when oauth is used Previously url_helper.py would raise exception if no oauth module was present. That is definitely reasonable. However, in the jenkins 'vmtest' test environment there was no oauth available. Everything else would have been fine, and only 'curtin pack' is executed from the host. It feels like curtin pack is sane to want to run with minimal dependencies. I do definitely want to be careful with the dynamic manner here. However, outside of MAAS reporter using oauth, it really is a unlikely path to be hit. And MAAS's use case is still satisfied because --install-deps is provided.	Scott Moser	8 years ago
330	block_meta: handle 'preserve' flag for raid devices block_meta: handle 'preserve' flag for raid devices The raid handler did not respect the 'preserve' flag and attempted to recreate the raid device instead of verifying it. Migrate mdadm usage into curtin.block.mdadm module. Complete mdadm unittest for new module.	Scott Moser	8 years ago
329	close file descriptors from --config= arguments close file descriptors from --config= arguments when running lvm installs, lvm over-zealously says things like: File descriptor 3 (configs/config-000.cfg) leaked ... This is described for grub at 1313784 and a workaround there too. The one listed here actually is a leaked file descriptor though. Those are opened. After consuming we should close.	Scott Moser	8 years ago
328	xenial: disable update-motd during an apt-get update xenial: disable update-motd during an apt-get update as reported in bug 1527710, xenial is starting a background task after 'apt-get update' is run. That was causing our unmount of the target to fail and fail the installation. The fix here is to disable the execution of that file and then restore it if we did so.	Scott Moser	8 years ago
327	re-work some of the test framework for less copying. re-work some of the test framework for less copying. here we do the following cleanup: a.) re-order some of the tests in a file, so that for a given test, we do in release order (Precice, Trusty, Vivid..) this is purely cosmetic and not enforced by anything. b.) add releases.py and use it to define the what was being defined in each test case. That removes duplication of 'arch', 'repo', and 'release' c.) generally clean up duplication that was present d.) use __test__ = False for any class that is a subclass of TestCase that should not be tested. That makes nosetests skip it.	Scott Moser	8 years ago

Older »