This should fix#435 for Helmfile v0.x releases since the next v0.150.0.
We introduce a new envvar to opt-in to the new YAML library, so that you can give it a shot before upgrading your Helmfile to v1. The same envvar can be used to opt-out of the new YAML library after you upgrade to Helmfile v1, giving you a more flexible migration story.
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
Previously a values template rendering failure resulted in a cryptic error message like:
```
g release templates in "input.yaml": failed executing templates in release "input.yaml"."foo1": failed executing template expressions in release "foo1".values[0] = "[98 97 115 101 76 97 98 101 108 58 32 34 123 123 32 103 101 116 32 46 82 101 108 101 97 115 101 46 76 97 98 101 108 115 32 92 34 98 97 115 101 92 34 32 92 34 78 79 84 95 73 78 72 69 82 73 84 69 68 92 34 32 125 125 34 10]": template: stringTemplate:1: unexpected "\\" in operand
```
The issue was that the values template at the value position of the YAML dict was printed as a formatted byte array. We now print it as a standard string so that it is easily readable.
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
* fix env value lost in environment values
Signed-off-by: yxxhero <aiopsclub@163.com>
* add more test
Signed-off-by: yxxhero <aiopsclub@163.com>
Signed-off-by: yxxhero <aiopsclub@163.com>
This is a successor to #596. We need a smooth migration path from `gopkg.in/yaml.v2`, and this pull request moves it forward with `goccy/go-yaml` instead of `gopkg.in/yaml.v3`. Merging this unblocks users stuck in Helmfile v0.146.x or earlier due to #435, so that they can upgrade to 0.147.x or greater without updating their helmfile configs.
We previously tried to upgrade to `yaml.v3` (https://github.com/helmfile/helmfile/issues/394) in Helmfile v0.x, presuming it won't break anything. Apparently, it broke use-cases where you want to layer release's `values` field over three or more release templates and releases (#435).
We then tried to bring back `yaml.v2` for Helmfile v0.x and keep `yaml.v3` for the upcoming Helmfile v1. However, it failed due to incompatibility in the Unmarshaller interface between `yaml.v2` and `yaml.v3` (https://github.com/helmfile/helmfile/pull/596).
`goccy/go-yaml` is, from my observation, a well-maintained alternative to `yaml.v2`. One of its premises is that it enables us to swap the implementation from `gopkg.in/yaml.v2` to `goccy/go-yaml` just by replacing the import directive. It seems to use the same `Unmarshaller` interface as yaml.v2 too.
Once this PR gets merged, I'd like to follow-up with adding a new build-time variable and an envvar to set the proper default for the yaml parser Helmfile uses and the ability to switch the parser at runtime. All in all, the next Helmfile release, v0.150.0 will get reverted to use `gopkg.in/yaml.v2` by default which resolves#435.
New users who started using Helmfile since any of v0.148.0, v0.148.1, and v0.149.0 might be already relying on the new behavior, They might need to specify a new envvar to enable `goccy/go-yaml`.
Signed-off-by: yxxhero <aiopsclub@163.com>
Signed-off-by: yxxhero <aiopsclub@163.com>
Co-authored-by: yxxhero <aiopsclub@163.com>
This is a successor to #440 rebuilt on top of #594 so that we can merge this while we are still at Hemlfile v0.x without worrying any backward-incompatibility. Much appreciation to @yxxhero for the original work!
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
* feat: Helmfile V1 mode
We add a new "V1 mode" to Helmfile so that you can seemlessly upgrade Helmfile from the current v0.x to the upcoming v1.0.
The idea is that we build both v0 and v1 binaries from the same tagged commit within the main branch, with different defaults for the "V1 mode"- the V1 mode is disabled by default for v0.x binaries, while it is enabled by default for v1.x binaries.
The V1 mode can be overrode at runtime via envvar. That is, even after upgrading the binary to v1, you will not see any backward-incompatible changes while you explicitly set an envvar, `HELMFILE_V1MODE=true`, at runtime.
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
Signed-off-by: yxxhero <aiopsclub@163.com>
All the dependencies get correctly installed when dealing with remote
charts.
If there's a local chart that depends on remote dependencies then those
don't get automatically installed. See #526. They end up with this
error:
```
Error: no cached repository for helm-manager-b6cf96b91af4f01317d185adfbe32610179e5246214be9646a52cb0b86032272 found. (try 'helm repo update'): open /root/.cache/helm/repository/helm-manager-b6cf96b91af4f01317d185adfbe32610179e5246214be9646a52cb0b86032272-index.yaml: no such file or directory
```
One workaround for that would be to add the repositories from the local
charts. Something like this:
```
cd local-chart/ && helm dependency list $dir 2> /dev/null | tail +2 | head -n -1 | awk '{ print "helm repo add " $1 " " $3 }' | while read cmd; do $cmd; done
```
This however is not trivial to parse and implement.
An easier fix which I did here is just to not allow doing
`--skip-refresh` for local repositories.
Fixes#526
Signed-off-by: Indrek Juhkam <indrek@urgas.eu>
Signed-off-by: Indrek Juhkam <indrek@urgas.eu>
Signed-off-by: yxxhero <aiopsclub@163.com>
1. only implement post-renderer flags this patch
2. As mumoshu advise, add helmfile flags `--post-render` and add the
postRenderer config in helmDefaults and release. the priority is
helmfile flags > release > helmDefaults.
3. fix the test case in state_test.go and some other tests.
Signed-off-by: guofutan <guofutan@tencent.com>
Signed-off-by: yxxhero <aiopsclub@163.com>
* Fix presync hooks are not called on no diff when run apply subcommand
Signed-off-by: xiaomudk <xiaomudk@gmail.com>
* Update docs/index.md
Co-authored-by: Yusuke Kuoka <ykuoka@gmail.com>
Signed-off-by: xiaomudk <xiaomudk@gmail.com>
Signed-off-by: xiaomudk <xiaomudk@gmail.com>
Co-authored-by: Yusuke Kuoka <ykuoka@gmail.com>
Use the new runWithLogCapture helper instead of the long boilerplate to capture the log for snapshot testing.
Signed-off-by: Yusuke Kuoka <ykuoka@gmail.com>
Allow configuring the lockfile in the state. This makes it possible for
example maintain a lock per environment.
Signed-off-by: Lassi Pölönen <lassi.polonen@iki.fi>
Signed-off-by: Lassi Pölönen <lassi.polonen@iki.fi>
This improves the `helmfile sync` performance.
From the code: `BuildDeps` is used only by `runHelmDepBuilds`, which
only is used by `PrepareCharts` which is finally only used by
`withPreparedCharts`.
`withPreparedCharts` already does `SyncReposOnce` which means we do not
have to refresh the local repository cache on each chart build.
This is only supported in Helm v3.
This seems to be mostly affecting helmfiles which have a lot of releases
and those release charts use sub dependencies.
I saw significant performance improvements for a helmfile with 45
releases, 2 repositories, and most of the charts also had their own
dependencies. Results:
Before the patch:
* real 9m10.565s
* real 9m38.335s
* real 9m14.941s
* real 5m13.106s (with cache)
After the patch:
* real 6m51.965s
* real 6m36.605s
* real 6m31.685s
* real 3m0.271s (with cache)
These were tested with:
```
rm -rf ~/.cache/helmfile ~/.cache/helm ~/.config/helm/repositories.* && helmfile sync ...
```
The result with `(with cache)` was without deleting the caches first.
From these metrics it seems that the sync duration decreased 20-45%
depending on the run, release count, dependencies and if the cache was
used or not.
As far as I understand, this should be backward-compatible change.
Signed-off-by: Indrek Juhkam <indrek@urgas.eu>
Signed-off-by: Indrek Juhkam <indrek@urgas.eu>
yxxhero
Yusuke Kuoka
Quan TRAN
Indrek Juhkam
guofutan
xiaomudk
koralsky
yxxhero
Lassi Pölönen
Helge Eichelberg