fix: stop artifact uploads depending on the cache server reaching Gitea (#1216)

Cache v2 makes the cache server the `ACTIONS_RESULTS_URL` origin, so artifact calls arrived there and were proxied on to Gitea, failing whenever it could not reach the instance.

- Artifact calls are answered with a redirect, so the cache server opens no connection to Gitea. A scheme change or an untrusted instance is still proxied, but there the cache server is the runner itself, which already reaches Gitea.
- Failures answer in twirp, not an empty `502` that clients report as `Unexpected end of JSON input`.
- `cache.v2: false` really points artifacts at Gitea now.
- Cache reservations are bound to the job that made them, so two jobs saving one key cannot commit against each other's upload, and a retry after a lost answer no longer fails a saved entry.
- The toolkit patch, which edits the GitHub-host check out of an action's bundle, was left in the shared checkout where a job running with `runner.patch_actions: false` could inherit it. It is put back after the job's copy.
- `exec` names an origin for the cache v2 it advertises, and masks its runtime token.

Behaviour changes: `no_proxy` no longer exempts `cache.external_server`, and `cache.enabled: false` also stops external registration.

Fixes https://gitea.com/gitea/runner/issues/1208
Fixes https://gitea.com/gitea/runner/issues/1211

Assisted by Claude (Opus 5).

Reviewed-on: https://gitea.com/gitea/runner/pulls/1216
Reviewed-by: bircni <bircni@icloud.com>
Co-authored-by: silverwind <me@silverwind.io>
This commit is contained in:
silverwind
2026-09-08 18:45:10 +00:00
committed by bircni
parent ff9965e940
commit 2ed8cdb76e
22 changed files with 533 additions and 82 deletions
+47 -11
View File
@@ -11,6 +11,7 @@ import (
"fmt"
"maps"
"net/http"
"net/url"
"os"
"path/filepath"
"runtime"
@@ -89,7 +90,7 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
envs := make(map[string]string, len(cfg.Runner.Envs))
maps.Copy(envs, cfg.Runner.Envs)
var cacheHandler *artifactcache.Handler
if cfg.Cache.Enabled == nil || *cfg.Cache.Enabled {
if cacheEnabled(cfg) {
if cfg.Cache.ExternalServer != "" {
warnIgnoredCachePolicy(cfg)
// The v1 client appends its path to this without a separator, so the slash is required.
@@ -109,6 +110,7 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
} else {
cacheHandler = handler
envs["ACTIONS_CACHE_URL"] = handler.ExternalURL() + "/"
warnIfCacheUnreachable(cfg, handler.ExternalURL())
}
}
}
@@ -388,7 +390,7 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
// Added per task because this job's service containers must be reached directly, and
// act reaches them by their workflow key.
proxyEnv := JobProxyEnv(envs, envs["ACTIONS_CACHE_URL"], slices.Sorted(maps.Keys(job.Services)))
proxyEnv := JobProxyEnv(envs, r.builtInCacheURL(), slices.Sorted(maps.Keys(job.Services)))
maps.Copy(envs, proxyEnv)
if r.capabilities != "" {
@@ -454,14 +456,7 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
// is that server's responsibility to authenticate requests.
revokeCache, resultsURL := r.registerCacheForTask(giteaRuntimeToken, preset.Repository, reporter)
defer revokeCache()
// A cache server that agreed to forward the artifact half is the whole results service, so the
// job is pointed at it.
if resultsURL != "" {
envs["ACTIONS_RESULTS_URL"] = resultsURL
if r.cacheServiceV2() {
envs[runner.CacheServiceV2Env] = "true"
}
}
r.setResultsService(envs, resultsURL)
eventJSON, err := json.Marshal(preset.Event)
if err != nil {
@@ -584,12 +579,43 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
return execErr
}
func (r *Runner) builtInCacheURL() string {
if r.cacheHandler == nil {
return ""
}
return r.envs["ACTIONS_CACHE_URL"]
}
func cacheEnabled(cfg *config.Config) bool {
return cfg.Cache.Enabled == nil || *cfg.Cache.Enabled
}
// cacheServiceV2 reports whether jobs are told the cache service speaks v2. It is all cache.v2
// turns off: the bundle edit that reaches it is what the artifact actions need too.
func (r *Runner) cacheServiceV2() bool {
return r.cfg.Cache.V2 == nil || *r.cfg.Cache.V2
}
func (r *Runner) setResultsService(envs map[string]string, resultsURL string) {
v2 := resultsURL != "" && r.cacheServiceV2()
if v2 {
envs[runner.CacheServiceV2Env] = "true"
} else {
delete(envs, runner.CacheServiceV2Env)
}
// Cache v2 clients read no other variable for it, and some cannot use the instance address.
if v2 || (resultsURL != "" && r.instanceOutOfReach(envs["ACTIONS_RESULTS_URL"])) {
envs["ACTIONS_RESULTS_URL"] = resultsURL
}
}
// These clients keep only the origin they are handed, and reach it on the job's own trust.
func (r *Runner) instanceOutOfReach(instance string) bool {
parsed, err := url.Parse(instance)
return err != nil || strings.Trim(parsed.Path, "/") != "" ||
(r.cfg.Runner.Insecure && parsed.Scheme == "https")
}
// registerCacheForTask tells the cache server to accept requests authenticated
// with the given runtime token for the duration of this task. Returns a
// function the caller must invoke (typically via defer) to revoke the
@@ -616,7 +642,7 @@ func (r *Runner) registerCacheForTask(token, repo string, reporter *report.Repor
if r.cacheHandler != nil {
return r.cacheHandler.RegisterJob(token, cred), r.cacheHandler.ResultsURL(cred)
}
if r.cfg.Cache.ExternalServer != "" && r.cfg.Cache.ExternalSecret != "" {
if cacheEnabled(r.cfg) && r.cfg.Cache.ExternalServer != "" && r.cfg.Cache.ExternalSecret != "" {
return r.registerExternalCacheJob(token, cred, reporter)
}
// No cache server to register against: caching is disabled, or the built-in server failed to start.
@@ -799,3 +825,13 @@ func warnIgnoredCacheSecret(cfg *config.Config) {
}
log.Warnf("%s is set but cache.external_server is not; the built-in cache server does not use a shared secret, so the value is ignored", key)
}
func warnIfCacheUnreachable(cfg *config.Config, cacheURL string) {
if cfg.Cache.Host != "" || cfg.Container.Network != "" {
return
}
if _, err := os.Stat("/.dockerenv"); err != nil {
return
}
log.Warnf("jobs are given %s for the cache server; if they cannot reach it, set container.network to a network this runner is on, or cache.host", cacheURL)
}
+7
View File
@@ -194,6 +194,13 @@ func TestRunner_ExternalCacheServer_RegisterRevoke(t *testing.T) {
require.Equal(t, http.StatusUnauthorized, probe(),
"token must be unknown to the remote server before registration")
disabled := &Runner{cfg: &config.Config{Cache: config.Cache{
Enabled: new(bool), ExternalServer: external, ExternalSecret: secret,
}}}
disabled.registerCacheForTask(token, repo, nil)
require.Equal(t, http.StatusUnauthorized, probe(),
"a disabled cache registers nothing, whatever external server is left configured")
unregister, resultsURL := r.registerCacheForTask(token, repo, nil)
require.NotEqual(t, http.StatusUnauthorized, probe(),
"token must be accepted after registerCacheForTask")
+31 -3
View File
@@ -143,6 +143,7 @@ func TestNewRunnerLeavesProxyToTheTask(t *testing.T) {
require.NotContains(t, r.envs, "http_proxy")
require.NotContains(t, r.envs, "no_proxy")
assert.Empty(t, r.builtInCacheURL(), "an external cache server is the operator's to exempt, not ours")
}
func taskWithDefaultActionsURL(url string) *runnerv1.Task {
@@ -189,11 +190,38 @@ func TestNewRunnerCacheServiceV2(t *testing.T) {
defer resp.Body.Close()
assert.Equal(t, http.StatusOK, resp.StatusCode, "the advertised results service serves no cache service")
assert.Equal(t, r.envs["ACTIONS_CACHE_URL"], r.builtInCacheURL(), "the address only the runner knows is bypassed for the operator")
envs := r.cloneEnvs()
r.setResultsService(envs, resultsURL)
assert.Equal(t, resultsURL, envs["ACTIONS_RESULTS_URL"])
assert.Equal(t, "true", envs[runner.CacheServiceV2Env])
// Turning v2 off withdraws the advertisement and nothing else.
assert.True(t, r.cacheServiceV2())
cfg.Cache.V2 = new(bool)
assert.False(t, r.cacheServiceV2())
envs = r.cloneEnvs()
envs[runner.CacheServiceV2Env] = "true"
r.setResultsService(envs, resultsURL)
assert.Equal(t, "https://gitea.example", envs["ACTIONS_RESULTS_URL"])
assert.Empty(t, envs[runner.CacheServiceV2Env], "a runner.envs entry would promise v2 at an origin not serving it")
envs = r.cloneEnvs()
envs[runner.CacheServiceV2Env] = "true"
envs["ACTIONS_RESULTS_URL"] = "https://gitea.example/sub"
r.setResultsService(envs, "")
assert.Equal(t, "https://gitea.example/sub", envs["ACTIONS_RESULTS_URL"], "with no cache server there is nothing to front with")
assert.Empty(t, envs[runner.CacheServiceV2Env])
for instance, insecure := range map[string]bool{
"https://gitea.example/sub": false,
"https://self-signed.example": true,
} {
cfg.Runner.Insecure = insecure
envs = r.cloneEnvs()
envs["ACTIONS_RESULTS_URL"] = instance
r.setResultsService(envs, resultsURL)
assert.Equal(t, resultsURL, envs["ACTIONS_RESULTS_URL"], instance)
assert.Empty(t, envs[runner.CacheServiceV2Env])
}
}
// The v1 cache client appends its path to ACTIONS_CACHE_URL without a separator, so a configured