There are some places where releases are only located if they are in the
state DEPLOYED. That particular logic was incorrectly used for upgrades.
That caused #1566. While fixing that issue, I found that this was also
the root cause of #1587 (though because it was off by one). I added a
generic method to get the last release, regardless of its status.
This allows some behaviors that previously failed:
- 'helm upgrade' can now be performed on a DELETED release
- 'helm rollback' can now be performed on a DELETED release even if
there is only one revision of that release history.
Closes#1566Closes#1587
This adds a hidden 'helm doc' command that generates documentation out
of the Helm help text.
Currently it can create:
- man pages for HELM(1)
- Markdown documentation
- bash autocompletions
Those are the three built-ins for Cobra.
The command is hidden so that it doesn't show up to the regular user,
since this is really a specialized task. It has the advantage of making
it trivially easy to install man pages and bash completion at installation time.
There were several places where an invalid name could be interpreted
into a wild-card by Kubernetes. That allows Bad Things.
This fix requires names to match the Kubernetes pattern for naming.
Closes#1594
Adds validation against the swagger schema.
Example error:
Error: release telling-wildebeest failed: error validating "": error
validating data: expected type int, for field
spec.template.spec.containers[0].ports[0].containerPort, got string
Current error:
unable to decode "": [pos 177]: json: expect char '"' but got char 'n'"'
This does the following to improve deletion failure handling:
- In an UninstallRelease operation, the release is marked DELETED
as soon as the basic checks are passed. This resolves 1508. I filed a
followup issue for doing this even better when we can modify protos
again.
- If a YAML manifest fails to parse, the error messages now suggests
that the record is corrupt, and the resources must be manually
deleted.
- If a resource is missing during deletion, the error messages now make
it clear that an object was skipped, but that the deletion continued.
Closes#1508
Long ago, Helm did not support cross-namespace installs. There was a
linter rule to catch this. When we changed the way Helm worked, we did
not remove the linter rule. This commit removes that linter rule.
Closes#1489
For archived files the Chart.yaml file should be contained in a base
directory. This commit adds an error when the Chart.yaml file is found
in the root directory.
Fixes#1171
This feature has been disabled in the past because simply enabling the
feature causes the Kubernetes library to make requests to a server.
Thus, running any tests that use the 'pkg/kube' library has required
running a kube API server.
This patch makes it possible to selectively activate 3rdParty API
support, and then disables that support during testing.
Future versions of the Kubernetes library appear to make it easier to
configure and mock this behavior, so this is most likely a stop-gap
measure. (Famous last words.)
Closes#1472
This switches the local chart repo to use 127.0.0.1 instead of localhost
so that the net library does not negotiate localhost to an IPv6 address,
which is known to cause issues in some Docker containers.
Breaking Change: When testing on a system that does NOT have IPv4
configured, this will break 'helm serve'. We estimate this will impact
none of the current user base.
Closes#1410
This commit changes the configmap storage driver to compress
the serialized release before storing it as a base64 encoded string.
This change is backward compatible as it handles existing releases
gracefully by skipping the decompression step when the gzip magic
header is not present.
Previous versions of Helm had placeholder text in the index.yaml file.
This generates an HTML index for 'helm serve'. It also has a
refactoring of the server so that the server can be tested.
Closes#1397
This prevents the index command from recursing through directories.
Behind the scenes, it swaps out the repository logic for the index file
logic.
Closes#1328
This adds a function engine.FuncMap that returns a function mapping that
better represents the functions passed to a template. The linting logic
is reconfigured to use this function instead of the sprig.FuncMap
function.
Closes#1366
The YAML validation was broken because the renderedContentMap has keys
with the path to the template including the chart directory, whereas the
linter was trying to access it relative from the chart directory.
For example, the hash key was `drupal/templates/deployment.yaml` but the
linter was trying to access `templates/deployment.yaml`. This commit
fixes the key used to access the rendered content.
Between Alpha.4 and Alpha.5 there was a change in the indexing logic.
This prevent indices from being appended to (because those index files
were often broken). This change allows the user to explicitly merge an
existing index and a generated index.
Closes#1334
- decouple tunnel from kube client
- add context switching for init cmd
- add unit tests for installer and init command
- refactor installer and remove unused code
This causes 'helm dep [up|install]' to ignore files in charts/ that
start with either a dot or an underscore. It also changes the
chartloader to ignore those files.
Also, if a 'helm dep up' does not find a charts/ directory, it creates
one.
Closes#1342
I wrote urlJoin to fix URL joining, but I put it in the wrong place and
never used it. This moves it to the right place, and replaces the hacky
previous solution.
This removes the requirement that a fetch or install command must
explicitly state the version number to install. Instead, this goes to
the strategy used by OS package managers: Install the latest until told
to do otherwise.
Closes#1198
This performs a relatively weak in-memory translation of index file
data. It does not, in most cases, write the corrected data to disk, and
it emits a warning directly to STDERR each time it loads a deprecated
index.
Known limitations:
- It cannot recover certain bogus records that earlier alpha releases
generated (notably, where all chartfile data is missing)
- In some cases, it has to parse a filename to get version info. This is
lossy.
- Because it takes three passes through the YAML and JSON unmarshal, it
is not performant.
This feature is transitional and should be removed during the Beta
cycle, prior to the release of 2.0.0.
Closes#1265
When `helm install -f foo.yaml bar` is called, and then the release is
upgraded with `helm upgrade happy-panda bar`, this will now re-use the
values that were submitted with `-f foo.yaml`. The same is true for
values specified with `--set`.
Closes#1227
There was a bug in the repo tests that caused them to overwrite the
repositories.yaml file in that directory. Now, the entire tests (server
and client-side) run inside of a temp directory.
This implements a new index file format for repository indices. It also
implements a new format for requirements.yaml.
Breaking change: This will break all previous versions of Helm, and will
impact helm search, repo, serve, and fetch functions.
Closes#1197
Some fields were introduced to Chart metadata after Alpha.4, but are not
going to be used ever. So we made the decision to remove them before we
get stuck with ugliness.
This should not break compatibility, since these fields were not used.
This also refactors significant portions of the CLI, moving much of the
shared code into a library.
Also in this release, a testing repository server has been added.
This is a minor refactor to move a leftover from Ancient Helm into the
current design. Specifically, the code to install Tiller from the Helm
client is now in `cmd/helm/installer`.
Closes#1033
This adds support for the following 'helm list' flags:
- all: show all release types
- deleted: show deleted releases
- deployed: show deployed releases
- failed: show failed releases
These flags can be toggled. Only '--deployed' is turned on by default.
On the server side, Tiller's list function can now filter based on a
slice of release.Status_Code filters. While the client only supports a
subset, the server supports all known release.Status_Code types.
Closes#973
Currently, delete returns an error if a resource is not found.
That is a bit confusing because if you delete something that
is not found, the end state is not really an error. The end
state is what you wanted to do in the first place. This type
of error is not propogated from the kube client but it is logged
in tiller. We could also change this to keep returning errors
(even the not found) and filter it out in the release server.
This adds the --verify and --keyring flags to:
helm fetch
helm inspect
helm install
helm upgrade
Each of these commands can now make cryptographic verification a
prerequisite for using a chart.
This adds support for OpenPGP signatures containing provenance data.
Such information can be used to verify the integrity of a Chart by
testing that its file hash, metadata, and images are correct.
This first PR does not contain all of the tooling necessary for
end-to-end chart integrity. It contains just the library.
See #983