"net/http"
"net/http/httptest"
"regexp"
+ "runtime"
"strings"
"time"
+ "git.arvados.org/arvados.git/lib/cmd"
"git.arvados.org/arvados.git/lib/config"
"git.arvados.org/arvados.git/sdk/go/arvados"
"git.arvados.org/arvados.git/sdk/go/arvadostest"
cluster.SystemRootToken = arvadostest.SystemRootToken
cluster.Collections.BlobSigningKey = arvadostest.BlobSigningKey
cluster.Volumes["z"] = arvados.Volume{StorageClasses: map[string]bool{"default": true}}
+ cluster.Containers.LocalKeepBlobBuffersPerVCPU = 0
s.handler = &Aggregator{Cluster: cluster}
s.req = httptest.NewRequest("GET", "/_health/all", nil)
s.req.Header.Set("Authorization", "Bearer "+arvadostest.ManagementToken)
s.checkOK(c)
}
+func (s *AggregatorSuite) TestClockSkew(c *check.C) {
+ // srv1: report real wall clock time
+ handler1 := healthyHandler{}
+ srv1, listen1 := s.stubServer(&handler1)
+ defer srv1.Close()
+ // srv2: report near-future time
+ handler2 := healthyHandler{headerDate: time.Now().Add(3 * time.Second)}
+ srv2, listen2 := s.stubServer(&handler2)
+ defer srv2.Close()
+ // srv3: report far-future time
+ handler3 := healthyHandler{headerDate: time.Now().Add(3*time.Minute + 3*time.Second)}
+ srv3, listen3 := s.stubServer(&handler3)
+ defer srv3.Close()
+
+ s.setAllServiceURLs(listen1)
+
+ // near-future time => OK
+ s.resp = httptest.NewRecorder()
+ arvadostest.SetServiceURL(&s.handler.Cluster.Services.DispatchCloud,
+ "http://localhost"+listen2+"/")
+ s.handler.ServeHTTP(s.resp, s.req)
+ s.checkOK(c)
+
+ // far-future time => error
+ s.resp = httptest.NewRecorder()
+ arvadostest.SetServiceURL(&s.handler.Cluster.Services.WebDAV,
+ "http://localhost"+listen3+"/")
+ s.handler.ServeHTTP(s.resp, s.req)
+ resp := s.checkUnhealthy(c)
+ if c.Check(len(resp.Errors) > 0, check.Equals, true) {
+ c.Check(resp.Errors[0], check.Matches, `clock skew detected: maximum timestamp spread is 3m.* \(exceeds warning threshold of 1m\)`)
+ }
+}
+
+func (s *AggregatorSuite) TestVersionSkew(c *check.C) {
+ // srv1: report same version
+ handler1 := healthyHandler{version: cmd.Version.String()}
+ srv1, listen1 := s.stubServer(&handler1)
+ defer srv1.Close()
+ // srv2: report same version but without " (go1.2.3)" part
+ handler2 := healthyHandler{version: strings.Fields(cmd.Version.String())[0]}
+ srv2, listen2 := s.stubServer(&handler2)
+ defer srv2.Close()
+ // srv3: report different version
+ handler3 := healthyHandler{version: "1.2.3~4 (" + runtime.Version() + ")"}
+ srv3, listen3 := s.stubServer(&handler3)
+ defer srv3.Close()
+
+ s.setAllServiceURLs(listen1)
+
+ // same version but without go1.2.3 part => OK
+ s.resp = httptest.NewRecorder()
+ arvadostest.SetServiceURL(&s.handler.Cluster.Services.RailsAPI,
+ "http://localhost"+listen2+"/")
+ s.handler.ServeHTTP(s.resp, s.req)
+ s.checkOK(c)
+
+ // different version => error
+ s.resp = httptest.NewRecorder()
+ arvadostest.SetServiceURL(&s.handler.Cluster.Services.WebDAV,
+ "http://localhost"+listen3+"/")
+ s.handler.ServeHTTP(s.resp, s.req)
+ resp := s.checkUnhealthy(c)
+ if c.Check(len(resp.Errors) > 0, check.Equals, true) {
+ c.Check(resp.Errors[0], check.Matches, `version mismatch: \Qkeep-web+http://localhost`+listen3+`\E is running 1.2.3~4 (.*) -- expected \Q`+cmd.Version.String()+`\E`)
+ }
+}
+
func (s *AggregatorSuite) TestPingTimeout(c *check.C) {
s.handler.timeout = arvados.Duration(100 * time.Millisecond)
srv, listen := s.stubServer(&slowHandler{})
confdata = regexp.MustCompile(`Source(Timestamp|SHA256): [^\n]+\n`).ReplaceAll(confdata, []byte{})
err = ioutil.WriteFile(tmpdir+"/config.yml", confdata, 0777)
c.Assert(err, check.IsNil)
+
var stdout, stderr bytes.Buffer
+
exitcode := CheckCommand.RunCommand("check", []string{"-config=" + tmpdir + "/config.yml"}, &bytes.Buffer{}, &stdout, &stderr)
c.Check(exitcode, check.Equals, 0)
+ c.Check(stderr.String(), check.Equals, "health check OK\n")
+ c.Check(stdout.String(), check.Equals, "")
+
+ stdout.Reset()
+ stderr.Reset()
+ exitcode = CheckCommand.RunCommand("check", []string{"-quiet", "-config=" + tmpdir + "/config.yml"}, &bytes.Buffer{}, &stdout, &stderr)
+ c.Check(exitcode, check.Equals, 0)
+ c.Check(stderr.String(), check.Equals, "")
+ c.Check(stdout.String(), check.Equals, "")
+
+ stdout.Reset()
+ stderr.Reset()
+ exitcode = CheckCommand.RunCommand("check", []string{"-config=" + tmpdir + "/config.yml", "-yaml"}, &bytes.Buffer{}, &stdout, &stderr)
+ c.Check(exitcode, check.Equals, 0)
c.Check(stderr.String(), check.Equals, "")
- c.Check(stdout.String(), check.Matches, `(?ms).*(\n|^)health: OK\n.*`)
+ c.Check(stdout.String(), check.Matches, `(?ms).*(\n|^)Health: OK\n.*`)
}
func (s *AggregatorSuite) checkError(c *check.C) {
&svcs.Controller,
&svcs.DispatchCloud,
&svcs.DispatchLSF,
+ &svcs.DispatchSLURM,
+ &svcs.GitHTTP,
&svcs.Keepbalance,
&svcs.Keepproxy,
&svcs.Keepstore,
}
type healthyHandler struct {
+ version string
configHash string
configTime time.Time
+ headerDate time.Time
}
func (h *healthyHandler) ServeHTTP(resp http.ResponseWriter, req *http.Request) {
+ if !h.headerDate.IsZero() {
+ resp.Header().Set("Date", h.headerDate.Format(time.RFC1123))
+ }
authOK := req.Header.Get("Authorization") == "Bearer "+arvadostest.ManagementToken
if req.URL.Path == "/_health/ping" {
if !authOK {
# HELP arvados_config_source_timestamp_seconds Timestamp of config file when it was loaded.
# TYPE arvados_config_source_timestamp_seconds gauge
arvados_config_source_timestamp_seconds{sha256="%s"} %g
+# HELP arvados_version_running Indicated version is running.
+# TYPE arvados_version_running gauge
+arvados_version_running{version="%s"} 1
`,
h.configHash, float64(time.Now().UnixNano())/1e9,
- h.configHash, float64(t.UnixNano())/1e9)
+ h.configHash, float64(t.UnixNano())/1e9,
+ h.version)
} else {
http.Error(resp, "not found", http.StatusNotFound)
}