[ARVADOS] created: bd1580a11a377270a5a7eed5abc8a6dfe6f9547d
git at public.curoverse.com
git at public.curoverse.com
Tue Oct 6 13:55:45 EDT 2015
at bd1580a11a377270a5a7eed5abc8a6dfe6f9547d (commit)
commit bd1580a11a377270a5a7eed5abc8a6dfe6f9547d
Author: radhika <radhika at curoverse.com>
Date: Tue Oct 6 13:54:41 2015 -0400
7167: get index from src and dst and copy any missing blocks from src to dst.
diff --git a/tools/keep-rsync/keep-rsync.go b/tools/keep-rsync/keep-rsync.go
index 3761cc7..c06fb8a 100644
--- a/tools/keep-rsync/keep-rsync.go
+++ b/tools/keep-rsync/keep-rsync.go
@@ -1,6 +1,7 @@
package main
import (
+ "bytes"
"flag"
"git.curoverse.com/arvados.git/sdk/go/arvadosclient"
"git.curoverse.com/arvados.git/sdk/go/keepclient"
@@ -69,6 +70,7 @@ func main() {
var err error
+ // Load config
if srcConfigFile == "" {
log.Fatal("-src-config-file must be specified.")
}
@@ -85,10 +87,14 @@ func main() {
log.Fatal("Error reading destination configuration: %s", err.Error())
}
+ // Initialize keep-rsync
err = initializeKeepRsync()
if err != nil {
log.Fatal("Error configurating keep-rsync: %s", err.Error())
}
+
+ // Copy blocks not found in dst from src
+ performKeepRsync()
}
// Reads config from file
@@ -157,6 +163,123 @@ func initializeKeepRsync() (err error) {
return
}
}
+ kcDst.Want_replicas = replications
return
}
+
+// Get unique block locators from src and dst
+// Copy any blocks missing in dst
+func performKeepRsync() error {
+ // Get unique locators from src
+ srcIndex, err := getUniqueLocators(kcSrc, prefix)
+ if err != nil {
+ return err
+ }
+
+ // Get unique locators from dst
+ dstIndex, err := getUniqueLocators(kcDst, prefix)
+ if err != nil {
+ return err
+ }
+
+ // Get list of locators found in src, but missing in dst
+ toBeCopied := getMissingLocators(srcIndex, dstIndex)
+
+ // Copy each missing block to dst
+ copyBlocksToDst(toBeCopied)
+
+ return nil
+}
+
+// Get list of unique locators from the specified cluster
+func getUniqueLocators(kc *keepclient.KeepClient, prefix string) (map[string]bool, error) {
+ var indexBytes []byte
+
+ for uuid := range kc.LocalRoots() {
+ reader, err := kc.GetIndex(uuid, prefix)
+ if err != nil {
+ return nil, err
+ }
+
+ var readBytes []byte
+ readBytes, err = ioutil.ReadAll(reader)
+ if err != nil {
+ return nil, err
+ }
+
+ indexBytes = append(indexBytes, readBytes...)
+ }
+
+ // Got index; Now dedup it
+ locators := bytes.Split(indexBytes, []byte("\n"))
+
+ uniqueLocators := map[string]bool{}
+ for _, loc := range locators {
+ if len(loc) == 0 {
+ continue
+ }
+
+ locator := string(bytes.Split(loc, []byte(" "))[0])
+ if _, ok := uniqueLocators[locator]; !ok {
+ uniqueLocators[locator] = true
+ }
+ }
+ return uniqueLocators, nil
+}
+
+// Get list of locators that are in src but not in dst
+func getMissingLocators(srcLocators map[string]bool, dstLocators map[string]bool) []string {
+ var missingLocators []string
+ for locator := range srcLocators {
+ if _, ok := dstLocators[locator]; !ok {
+ missingLocators = append(missingLocators, locator)
+ }
+ }
+ return missingLocators
+}
+
+// Copy blocks from src to dst; only those that are missing in dst are copied
+func copyBlocksToDst(toBeCopied []string) {
+ done := 0
+ total := len(toBeCopied)
+ var failed []string
+
+ for _, locator := range toBeCopied {
+ log.Printf("Getting block %d of %d", done+1, total)
+
+ log.Printf("Getting block: %v", locator)
+
+ reader, _, _, err := kcSrc.Get(locator)
+ if err != nil {
+ log.Printf("Error getting block: %q %v", locator, err)
+ failed = append(failed, locator)
+ continue
+ }
+ data, err := ioutil.ReadAll(reader)
+ if err != nil {
+ log.Printf("Error reading block data: %q %v", locator, err)
+ failed = append(failed, locator)
+ continue
+ }
+
+ log.Printf("Copying block: %q", locator)
+ _, rep, err := kcDst.PutB(data)
+ if err != nil {
+ log.Printf("Error putting block data: %q %v", locator, err)
+ failed = append(failed, locator)
+ continue
+ }
+ if rep != replications {
+ log.Printf("Failed to put enough number of replicas. Wanted: %d; Put: %d", replications, rep)
+ failed = append(failed, locator)
+ continue
+ }
+
+ done++
+ log.Printf("%.2f%% done", float64(done)/float64(total)*100)
+ }
+
+ log.Printf("Successfully copied to destination %d and failed %d out of a total of %d", done, len(failed), total)
+ log.Printf("Failed blocks %v", failed)
+}
diff --git a/tools/keep-rsync/keep-rsync_test.go b/tools/keep-rsync/keep-rsync_test.go
index 97db571..37e4b20 100644
--- a/tools/keep-rsync/keep-rsync_test.go
+++ b/tools/keep-rsync/keep-rsync_test.go
@@ -55,6 +55,9 @@ func setupRsync(c *C) {
dstConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
dstConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
+ replications = 1
+
+ // Start API and Keep servers
arvadostest.StartAPI()
arvadostest.StartKeep()
@@ -150,7 +153,7 @@ func (s *ServerRequiredSuite) TestRsyncInitializeWithKeepServicesJSON(c *C) {
c.Check(localRoots != nil, Equals, true)
foundIt := false
- for k, _ := range localRoots {
+ for k := range localRoots {
if k == "zzzzz-bi6l4-123456789012340" {
foundIt = true
}
@@ -158,10 +161,93 @@ func (s *ServerRequiredSuite) TestRsyncInitializeWithKeepServicesJSON(c *C) {
c.Check(foundIt, Equals, true)
foundIt = false
- for k, _ := range localRoots {
+ for k := range localRoots {
if k == "zzzzz-bi6l4-123456789012341" {
foundIt = true
}
}
c.Check(foundIt, Equals, true)
}
+
+// Put 5 blocks in src. Put 2 of those blocks in dst
+// Hence there are 3 additional blocks in src
+// Also, put 2 extra blocks in dts; they are hence only in dst
+// Run rsync and verify that those 7 blocks are now available in dst
+func (s *ServerRequiredSuite) TestKeepRsync(c *C) {
+ setupRsync(c)
+
+ // Put a few blocks in src using kcSrc
+ var srcLocators []string
+ for i := 0; i < 5; i++ {
+ data := []byte(fmt.Sprintf("test-data-%d", i))
+ hash := fmt.Sprintf("%x", md5.Sum(data))
+
+ hash2, rep, err := kcSrc.PutB(data)
+ c.Check(hash2, Matches, fmt.Sprintf(`^%s\+11(\+.+)?$`, hash))
+ c.Check(rep, Equals, 2)
+ c.Check(err, Equals, nil)
+
+ reader, blocklen, _, err := kcSrc.Get(hash)
+ c.Assert(err, Equals, nil)
+ c.Check(blocklen, Equals, int64(11))
+ all, err := ioutil.ReadAll(reader)
+ c.Check(all, DeepEquals, data)
+
+ srcLocators = append(srcLocators, fmt.Sprintf("%s+%d", hash, blocklen))
+ }
+
+ // Put just two of those blocks in dst using kcDst
+ var dstLocators []string
+ for i := 0; i < 2; i++ {
+ data := []byte(fmt.Sprintf("test-data-%d", i))
+ hash := fmt.Sprintf("%x", md5.Sum(data))
+
+ hash2, rep, err := kcDst.PutB(data)
+ c.Check(hash2, Matches, fmt.Sprintf(`^%s\+11(\+.+)?$`, hash))
+ c.Check(rep, Equals, 1)
+ c.Check(err, Equals, nil)
+
+ reader, blocklen, _, err := kcDst.Get(hash)
+ c.Assert(err, Equals, nil)
+ c.Check(blocklen, Equals, int64(11))
+ all, err := ioutil.ReadAll(reader)
+ c.Check(all, DeepEquals, data)
+
+ dstLocators = append(dstLocators, fmt.Sprintf("%s+%d", hash, blocklen))
+ }
+
+ // Put two more blocks in dst; they are not in src at all
+ var extraDstLocators []string
+ for i := 0; i < 2; i++ {
+ data := []byte(fmt.Sprintf("other-data-%d", i))
+ hash := fmt.Sprintf("%x", md5.Sum(data))
+
+ hash2, rep, err := kcDst.PutB(data)
+ c.Check(hash2, Matches, fmt.Sprintf(`^%s\+12(\+.+)?$`, hash))
+ c.Check(rep, Equals, 1)
+ c.Check(err, Equals, nil)
+
+ reader, blocklen, _, err := kcDst.Get(hash)
+ c.Assert(err, Equals, nil)
+ c.Check(blocklen, Equals, int64(12))
+ all, err := ioutil.ReadAll(reader)
+ c.Check(all, DeepEquals, data)
+
+ extraDstLocators = append(extraDstLocators, fmt.Sprintf("%s+%d", hash, blocklen))
+ }
+
+ err := performKeepRsync()
+ c.Check(err, Equals, nil)
+
+ // Now GetIndex from dst and verify that all 5 from src and the 2 extra blocks are found
+ dstIndex, err := getUniqueLocators(kcDst, "")
+ c.Check(err, Equals, nil)
+ for _, locator := range srcLocators {
+ _, ok := dstIndex[locator]
+ c.Assert(ok, Equals, true)
+ }
+ for _, locator := range extraDstLocators {
+ _, ok := dstIndex[locator]
+ c.Assert(ok, Equals, true)
+ }
+}
commit 290247250e0c5702a341f98dfdc380e29f1b45c7
Author: radhika <radhika at curoverse.com>
Date: Mon Oct 5 17:22:37 2015 -0400
7167: Update test to also put a block in dst and attempt get from src.
diff --git a/sdk/python/tests/run_test_server.py b/sdk/python/tests/run_test_server.py
index 62a2a82..fba9bb6 100644
--- a/sdk/python/tests/run_test_server.py
+++ b/sdk/python/tests/run_test_server.py
@@ -353,9 +353,11 @@ def run_keep(blob_signing_key=None, enforce_permissions=False):
api.keep_disks().delete(uuid=d['uuid']).execute()
start_index = 0
+ end_index = 2
if keep_existing is not None:
start_index = 2
- for d in range(start_index, start_index+2):
+ end_index = 3
+ for d in range(start_index, end_index):
port = _start_keep(d, keep_args)
svc = api.keep_services().create(body={'keep_service': {
'uuid': 'zzzzz-bi6l4-keepdisk{:07d}'.format(d),
@@ -380,9 +382,8 @@ def _stop_keep(n):
def stop_keep():
_stop_keep(0)
_stop_keep(1)
- # We may have created 2 additional keep servers when keep_existing is used
+ # We may have created an additional keep servers when keep_existing is used
_stop_keep(2)
- _stop_keep(3)
def run_keep_proxy():
if 'ARVADOS_TEST_PROXY_SERVICES' in os.environ:
diff --git a/tools/keep-rsync/keep-rsync_test.go b/tools/keep-rsync/keep-rsync_test.go
index 7636c2e..97db571 100644
--- a/tools/keep-rsync/keep-rsync_test.go
+++ b/tools/keep-rsync/keep-rsync_test.go
@@ -68,6 +68,7 @@ func setupRsync(c *C) {
// load kcDst
kcDst, err = keepclient.MakeKeepClient(&arvDst)
c.Assert(err, Equals, nil)
+ kcDst.Want_replicas = 1
}
// Test readConfigFromFile method
@@ -94,27 +95,48 @@ func (s *ServerRequiredSuite) TestReadConfigFromFile(c *C) {
// Test keep-rsync initialization, with src and dst keep servers.
// Do a Put and Get in src, both of which should succeed.
-// Do a Get in dst for the same hash, which should raise block not found error.
-func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDstShouldFail(c *C) {
+// Do a Put and Get in dst, both of which should succeed.
+// Do a Get in dst for the src hash, which should raise block not found error.
+// Do a Get in src for the dst hash, which should raise block not found error.
+func (s *ServerRequiredSuite) TestRsyncPutInOne_GetFromOtherShouldFail(c *C) {
setupRsync(c)
// Put a block in src using kcSrc and Get it
- data := []byte("test-data")
- hash := fmt.Sprintf("%x", md5.Sum(data))
+ srcData := []byte("test-data1")
+ locatorInSrc := fmt.Sprintf("%x", md5.Sum(srcData))
- hash2, rep, err := kcSrc.PutB(data)
- c.Check(hash2, Matches, fmt.Sprintf(`^%s\+9(\+.+)?$`, hash))
+ hash, rep, err := kcSrc.PutB(srcData)
+ c.Check(hash, Matches, fmt.Sprintf(`^%s\+10(\+.+)?$`, locatorInSrc))
c.Check(rep, Equals, 2)
c.Check(err, Equals, nil)
- reader, blocklen, _, err := kcSrc.Get(hash)
+ reader, blocklen, _, err := kcSrc.Get(locatorInSrc)
c.Assert(err, Equals, nil)
- c.Check(blocklen, Equals, int64(9))
+ c.Check(blocklen, Equals, int64(10))
all, err := ioutil.ReadAll(reader)
- c.Check(all, DeepEquals, data)
+ c.Check(all, DeepEquals, srcData)
- // Get using kcDst should fail with NotFound error
- _, _, _, err = kcDst.Get(hash)
+ // Put a different block in src using kcSrc and Get it
+ dstData := []byte("test-data2")
+ locatorInDst := fmt.Sprintf("%x", md5.Sum(dstData))
+
+ hash, rep, err = kcDst.PutB(dstData)
+ c.Check(hash, Matches, fmt.Sprintf(`^%s\+10(\+.+)?$`, locatorInDst))
+ c.Check(rep, Equals, 1)
+ c.Check(err, Equals, nil)
+
+ reader, blocklen, _, err = kcDst.Get(locatorInDst)
+ c.Assert(err, Equals, nil)
+ c.Check(blocklen, Equals, int64(10))
+ all, err = ioutil.ReadAll(reader)
+ c.Check(all, DeepEquals, dstData)
+
+ // Get srcLocator using kcDst should fail with NotFound error
+ _, _, _, err = kcDst.Get(locatorInSrc)
+ c.Assert(err.Error(), Equals, "Block not found")
+
+ // Get dstLocator using kcSrc should fail with NotFound error
+ _, _, _, err = kcSrc.Get(locatorInDst)
c.Assert(err.Error(), Equals, "Block not found")
}
commit f3b5ffc2a4409d4c1b676e40ba4e582ca2beeb61
Author: radhika <radhika at curoverse.com>
Date: Mon Oct 5 11:41:39 2015 -0400
7167: Refactor MakeKeepClient and DiscoverKeepServers to allow making KeepClient from input JSON as well.
diff --git a/sdk/go/keepclient/keepclient.go b/sdk/go/keepclient/keepclient.go
index 53dfb2b..05fad6e 100644
--- a/sdk/go/keepclient/keepclient.go
+++ b/sdk/go/keepclient/keepclient.go
@@ -54,9 +54,20 @@ type KeepClient struct {
replicasPerService int
}
-// Create a new KeepClient. This will contact the API server to discover Keep
-// servers.
+// MakeKeepClient creates a new KeepClient by contacting the API server to discover Keep servers.
func MakeKeepClient(arv *arvadosclient.ArvadosClient) (*KeepClient, error) {
+ kc := initKeepClient(arv)
+ return kc, kc.DiscoverKeepServers()
+}
+
+// MakeKeepClientFromJSON creates a new KeepClient using the given json to load keep servers.
+func MakeKeepClientFromJSON(arv *arvadosclient.ArvadosClient, svcJSON string) (*KeepClient, error) {
+ kc := initKeepClient(arv)
+ return kc, kc.DiscoverKeepServersFromJSON(svcJSON)
+}
+
+// Make a new KeepClient struct.
+func initKeepClient(arv *arvadosclient.ArvadosClient) *KeepClient {
var matchTrue = regexp.MustCompile("^(?i:1|yes|true)$")
insecure := matchTrue.MatchString(os.Getenv("ARVADOS_API_HOST_INSECURE"))
kc := &KeepClient{
@@ -66,7 +77,7 @@ func MakeKeepClient(arv *arvadosclient.ArvadosClient) (*KeepClient, error) {
Client: &http.Client{Transport: &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: insecure}}},
}
- return kc, kc.DiscoverKeepServers()
+ return kc
}
// Put a block given the block hash, a reader, and the number of bytes
diff --git a/sdk/go/keepclient/support.go b/sdk/go/keepclient/support.go
index 63800b1..8be1780 100644
--- a/sdk/go/keepclient/support.go
+++ b/sdk/go/keepclient/support.go
@@ -2,6 +2,7 @@ package keepclient
import (
"crypto/md5"
+ "encoding/json"
"errors"
"fmt"
"git.curoverse.com/arvados.git/sdk/go/streamer"
@@ -76,19 +77,38 @@ func (this *KeepClient) setClientSettingsDisk() {
}
}
+type svcList struct {
+ Items []keepService `json:"items"`
+}
+
// DiscoverKeepServers gets list of available keep services from api server
func (this *KeepClient) DiscoverKeepServers() error {
- type svcList struct {
- Items []keepService `json:"items"`
- }
- var m svcList
+ var list svcList
// Get keep services from api server
- err := this.Arvados.Call("GET", "keep_services", "", "accessible", nil, &m)
+ err := this.Arvados.Call("GET", "keep_services", "", "accessible", nil, &list)
if err != nil {
return err
}
+ return this.loadKeepServers(list)
+}
+
+// DiscoverKeepServersFromJSON gets list of available keep services from given JSON
+func (this *KeepClient) DiscoverKeepServersFromJSON(services string) error {
+ var list svcList
+
+ // Load keep services from given json
+ dec := json.NewDecoder(strings.NewReader(services))
+ if err := dec.Decode(&list); err != nil {
+ return err
+ }
+
+ return this.loadKeepServers(list)
+}
+
+// loadKeepServers
+func (this *KeepClient) loadKeepServers(list svcList) error {
listed := make(map[string]bool)
localRoots := make(map[string]string)
gatewayRoots := make(map[string]string)
@@ -98,7 +118,7 @@ func (this *KeepClient) DiscoverKeepServers() error {
this.replicasPerService = 1
this.Using_proxy = false
- for _, service := range m.Items {
+ for _, service := range list.Items {
scheme := "http"
if service.SSL {
scheme = "https"
diff --git a/tools/keep-rsync/keep-rsync.go b/tools/keep-rsync/keep-rsync.go
index c9fd77a..3761cc7 100644
--- a/tools/keep-rsync/keep-rsync.go
+++ b/tools/keep-rsync/keep-rsync.go
@@ -120,22 +120,43 @@ var (
// Initializes keep-rsync using the config provided
func initializeKeepRsync() (err error) {
+ // arvSrc from srcConfig
arvSrc, err = arvadosclient.MakeArvadosClientWithConfig(srcConfig)
if err != nil {
return
}
+ // arvDst from dstConfig
arvDst, err = arvadosclient.MakeArvadosClientWithConfig(dstConfig)
if err != nil {
return
}
- kcSrc, err = keepclient.MakeKeepClient(&arvSrc)
- if err != nil {
- return
+ // if srcKeepServicesJSON is provided, use it to load services; else, use DiscoverKeepServers
+ if srcKeepServicesJSON == "" {
+ kcSrc, err = keepclient.MakeKeepClient(&arvSrc)
+ if err != nil {
+ return
+ }
+ } else {
+ kcSrc, err = keepclient.MakeKeepClientFromJSON(&arvSrc, srcKeepServicesJSON)
+ if err != nil {
+ return
+ }
}
- kcDst, err = keepclient.MakeKeepClient(&arvDst)
+ // if dstKeepServicesJSON is provided, use it to load services; else, use DiscoverKeepServers
+ if dstKeepServicesJSON == "" {
+ kcDst, err = keepclient.MakeKeepClient(&arvDst)
+ if err != nil {
+ return
+ }
+ } else {
+ kcDst, err = keepclient.MakeKeepClientFromJSON(&arvDst, dstKeepServicesJSON)
+ if err != nil {
+ return
+ }
+ }
return
}
diff --git a/tools/keep-rsync/keep-rsync_test.go b/tools/keep-rsync/keep-rsync_test.go
index c45fbf8..7636c2e 100644
--- a/tools/keep-rsync/keep-rsync_test.go
+++ b/tools/keep-rsync/keep-rsync_test.go
@@ -29,6 +29,8 @@ func (s *ServerRequiredSuite) SetUpSuite(c *C) {
func (s *ServerRequiredSuite) SetUpTest(c *C) {
arvadostest.ResetEnv()
+ srcKeepServicesJSON = ""
+ dstKeepServicesJSON = ""
}
func (s *ServerRequiredSuite) TearDownSuite(c *C) {
@@ -115,3 +117,29 @@ func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDstShouldFail(c *C) {
_, _, _, err = kcDst.Get(hash)
c.Assert(err.Error(), Equals, "Block not found")
}
+
+// Test keep-rsync initialization, with srcKeepServicesJSON
+func (s *ServerRequiredSuite) TestRsyncInitializeWithKeepServicesJSON(c *C) {
+ srcKeepServicesJSON = "{ \"kind\":\"arvados#keepServiceList\", \"etag\":\"\", \"self_link\":\"\", \"offset\":null, \"limit\":null, \"items\":[ { \"href\":\"/keep_services/zzzzz-bi6l4-123456789012340\", \"kind\":\"arvados#keepService\", \"etag\":\"641234567890enhj7hzx432e5\", \"uuid\":\"zzzzz-bi6l4-123456789012340\", \"owner_uuid\":\"zzzzz-tpzed-123456789012345\", \"service_host\":\"keep0.zzzzz.arvadosapi.com\", \"service_port\":25107, \"service_ssl_flag\":false, \"service_type\":\"disk\", \"read_only\":false }, { \"href\":\"/keep_services/zzzzz-bi6l4-123456789012341\", \"kind\":\"arvados#keepService\", \"etag\":\"641234567890enhj7hzx432e5\", \"uuid\":\"zzzzz-bi6l4-123456789012341\", \"owner_uuid\":\"zzzzz-tpzed-123456789012345\", \"service_host\":\"keep0.zzzzz.arvadosapi.com\", \"service_port\":25108, \"service_ssl_flag\":false, \"service_type\":\"disk\", \"read_only\":false } ], \"items_available\":2 }"
+
+ setupRsync(c)
+
+ localRoots := kcSrc.LocalRoots()
+ c.Check(localRoots != nil, Equals, true)
+
+ foundIt := false
+ for k, _ := range localRoots {
+ if k == "zzzzz-bi6l4-123456789012340" {
+ foundIt = true
+ }
+ }
+ c.Check(foundIt, Equals, true)
+
+ foundIt = false
+ for k, _ := range localRoots {
+ if k == "zzzzz-bi6l4-123456789012341" {
+ foundIt = true
+ }
+ }
+ c.Check(foundIt, Equals, true)
+}
commit 6823f2d65a48bc989b819b85b6428c230b223c82
Author: radhika <radhika at curoverse.com>
Date: Mon Oct 5 09:25:34 2015 -0400
7167: args not avaialble in all tests; hence store keep_existing argument in a variable rather than accessing it directly from args.
diff --git a/sdk/python/tests/run_test_server.py b/sdk/python/tests/run_test_server.py
index d5d1874..62a2a82 100644
--- a/sdk/python/tests/run_test_server.py
+++ b/sdk/python/tests/run_test_server.py
@@ -43,6 +43,7 @@ if not os.path.exists(TEST_TMPDIR):
my_api_host = None
_cached_config = {}
+keep_existing = None
def find_server_pid(PID_PATH, wait=10):
now = time.time()
@@ -324,7 +325,7 @@ def _start_keep(n, keep_args):
return port
def run_keep(blob_signing_key=None, enforce_permissions=False):
- if args.keep_existing is None:
+ if keep_existing is None:
stop_keep()
keep_args = {}
@@ -352,7 +353,7 @@ def run_keep(blob_signing_key=None, enforce_permissions=False):
api.keep_disks().delete(uuid=d['uuid']).execute()
start_index = 0
- if args.keep_existing is not None:
+ if keep_existing is not None:
start_index = 2
for d in range(start_index, start_index+2):
port = _start_keep(d, keep_args)
@@ -606,6 +607,8 @@ if __name__ == "__main__":
parser.add_argument('--keep_existing', type=str, help="Used to add additional keep servers, without terminating existing servers")
args = parser.parse_args()
+ keep_existing = args.keep_existing
+
if args.action not in actions:
print("Unrecognized action '{}'. Actions are: {}.".format(args.action, actions), file=sys.stderr)
sys.exit(1)
diff --git a/tools/keep-rsync/keep-rsync.go b/tools/keep-rsync/keep-rsync.go
index eff8b9c..c9fd77a 100644
--- a/tools/keep-rsync/keep-rsync.go
+++ b/tools/keep-rsync/keep-rsync.go
@@ -11,17 +11,17 @@ import (
// keep-rsync arguments
var (
- srcConfig map[string]string
- destConfig map[string]string
- srcKeepServicesJSON string
- destKeepServicesJSON string
- replications int
- prefix string
+ srcConfig map[string]string
+ dstConfig map[string]string
+ srcKeepServicesJSON string
+ dstKeepServicesJSON string
+ replications int
+ prefix string
)
func main() {
var srcConfigFile string
- var destConfigFile string
+ var dstConfigFile string
flag.StringVar(
&srcConfigFile,
@@ -32,8 +32,8 @@ func main() {
"ARVADOS_API_HOST, ARVADOS_API_HOST_INSECURE, and ARVADOS_BLOB_SIGNING_KEY.")
flag.StringVar(
- &destConfigFile,
- "dest-config-file",
+ &dstConfigFile,
+ "dst-config-file",
"",
"Destination configuration filename with full path that contains "+
"an ARVADOS_API_TOKEN which is a valid datamanager token recognized by the destination keep servers, "+
@@ -47,11 +47,11 @@ func main() {
"If not provided, this list is obtained from api server configured in src-config-file.")
flag.StringVar(
- &destKeepServicesJSON,
- "dest-keep-services-json",
+ &dstKeepServicesJSON,
+ "dst-keep-services-json",
"",
"An optional list of available destination keepservices. "+
- "If not provided, this list is obtained from api server configured in dest-config-file.")
+ "If not provided, this list is obtained from api server configured in dst-config-file.")
flag.IntVar(
&replications,
@@ -77,10 +77,10 @@ func main() {
log.Fatal("Error reading source configuration: %s", err.Error())
}
- if destConfigFile == "" {
- log.Fatal("-dest-config-file must be specified.")
+ if dstConfigFile == "" {
+ log.Fatal("-dst-config-file must be specified.")
}
- destConfig, err = readConfigFromFile(destConfigFile)
+ dstConfig, err = readConfigFromFile(dstConfigFile)
if err != nil {
log.Fatal("Error reading destination configuration: %s", err.Error())
}
@@ -112,10 +112,10 @@ func readConfigFromFile(filename string) (map[string]string, error) {
// keep-rsync source and destination clients
var (
- arvSrc arvadosclient.ArvadosClient
- arvDest arvadosclient.ArvadosClient
- kcSrc *keepclient.KeepClient
- kcDest *keepclient.KeepClient
+ arvSrc arvadosclient.ArvadosClient
+ arvDst arvadosclient.ArvadosClient
+ kcSrc *keepclient.KeepClient
+ kcDst *keepclient.KeepClient
)
// Initializes keep-rsync using the config provided
@@ -125,7 +125,7 @@ func initializeKeepRsync() (err error) {
return
}
- arvDest, err = arvadosclient.MakeArvadosClientWithConfig(destConfig)
+ arvDst, err = arvadosclient.MakeArvadosClientWithConfig(dstConfig)
if err != nil {
return
}
@@ -135,7 +135,7 @@ func initializeKeepRsync() (err error) {
return
}
- kcDest, err = keepclient.MakeKeepClient(&arvDest)
+ kcDst, err = keepclient.MakeKeepClient(&arvDst)
return
}
diff --git a/tools/keep-rsync/keep-rsync_test.go b/tools/keep-rsync/keep-rsync_test.go
index e2b1f0f..c45fbf8 100644
--- a/tools/keep-rsync/keep-rsync_test.go
+++ b/tools/keep-rsync/keep-rsync_test.go
@@ -36,7 +36,7 @@ func (s *ServerRequiredSuite) TearDownSuite(c *C) {
arvadostest.StopAPI()
}
-// Testing keep-rsync needs two sets of keep services: src and dest.
+// Testing keep-rsync needs two sets of keep services: src and dst.
// The test setup hence tweaks keep-rsync initialzation to achieve this.
// First invoke initializeKeepRsync and then invoke StartKeepAdditional
// to create the keep servers to be used as destination.
@@ -47,11 +47,11 @@ func setupRsync(c *C) {
srcConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
srcConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
- // destConfig
- destConfig = make(map[string]string)
- destConfig["ARVADOS_API_HOST"] = os.Getenv("ARVADOS_API_HOST")
- destConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
- destConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
+ // dstConfig
+ dstConfig = make(map[string]string)
+ dstConfig["ARVADOS_API_HOST"] = os.Getenv("ARVADOS_API_HOST")
+ dstConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
+ dstConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
arvadostest.StartAPI()
arvadostest.StartKeep()
@@ -63,8 +63,8 @@ func setupRsync(c *C) {
// Create two more keep servers to be used as destination
arvadostest.StartKeepAdditional(true)
- // load kcDest
- kcDest, err = keepclient.MakeKeepClient(&arvDest)
+ // load kcDst
+ kcDst, err = keepclient.MakeKeepClient(&arvDst)
c.Assert(err, Equals, nil)
}
@@ -90,10 +90,10 @@ func (s *ServerRequiredSuite) TestReadConfigFromFile(c *C) {
c.Assert(config["EXTERNAL_CLIENT"], Equals, "")
}
-// Test keep-rsync initialization, with src and dest keep servers.
+// Test keep-rsync initialization, with src and dst keep servers.
// Do a Put and Get in src, both of which should succeed.
-// Do a Get in dest for the same hash, which should raise block not found error.
-func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDestShouldFail(c *C) {
+// Do a Get in dst for the same hash, which should raise block not found error.
+func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDstShouldFail(c *C) {
setupRsync(c)
// Put a block in src using kcSrc and Get it
@@ -111,7 +111,7 @@ func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDestShouldFail(c *C) {
all, err := ioutil.ReadAll(reader)
c.Check(all, DeepEquals, data)
- // Get using kcDest should fail with NotFound error
- _, _, _, err = kcDest.Get(hash)
+ // Get using kcDst should fail with NotFound error
+ _, _, _, err = kcDst.Get(hash)
c.Assert(err.Error(), Equals, "Block not found")
}
commit cb48eb95d516b1d4a1ffe18be34703005531117a
Merge: 86df40e f81f84e
Author: radhika <radhika at curoverse.com>
Date: Mon Oct 5 07:48:47 2015 -0400
Merge branch 'master' into 7167-keep-rsync-test-setup
commit 86df40e33e586ccb4dc506e00f773392d454804c
Author: radhika <radhika at curoverse.com>
Date: Mon Oct 5 07:46:56 2015 -0400
7167: keep-rsync parameter loading and intialization. Update test framework to allow creating two sets of keep servers, source and destination.
diff --git a/sdk/go/arvadosclient/arvadosclient.go b/sdk/go/arvadosclient/arvadosclient.go
index 1cce0a7..ab2d9b2 100644
--- a/sdk/go/arvadosclient/arvadosclient.go
+++ b/sdk/go/arvadosclient/arvadosclient.go
@@ -82,13 +82,25 @@ type ArvadosClient struct {
// variables ARVADOS_API_HOST, ARVADOS_API_TOKEN, and (optionally)
// ARVADOS_API_HOST_INSECURE.
func MakeArvadosClient() (ac ArvadosClient, err error) {
+ config := make(map[string]string)
+ config["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
+ config["ARVADOS_API_HOST"] = os.Getenv("ARVADOS_API_HOST")
+ config["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
+ config["ARVADOS_EXTERNAL_CLIENT"] = os.Getenv("ARVADOS_EXTERNAL_CLIENT")
+
+ return MakeArvadosClientWithConfig(config)
+}
+
+// Create a new ArvadosClient, using the given input parameters.
+func MakeArvadosClientWithConfig(config map[string]string) (ac ArvadosClient, err error) {
var matchTrue = regexp.MustCompile("^(?i:1|yes|true)$")
- insecure := matchTrue.MatchString(os.Getenv("ARVADOS_API_HOST_INSECURE"))
- external := matchTrue.MatchString(os.Getenv("ARVADOS_EXTERNAL_CLIENT"))
+
+ insecure := matchTrue.MatchString(config["ARVADOS_API_HOST_INSECURE"])
+ external := matchTrue.MatchString(config["ARVADOS_EXTERNAL_CLIENT"])
ac = ArvadosClient{
- ApiServer: os.Getenv("ARVADOS_API_HOST"),
- ApiToken: os.Getenv("ARVADOS_API_TOKEN"),
+ ApiServer: config["ARVADOS_API_HOST"],
+ ApiToken: config["ARVADOS_API_TOKEN"],
ApiInsecure: insecure,
Client: &http.Client{Transport: &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: insecure}}},
diff --git a/sdk/go/arvadostest/run_servers.go b/sdk/go/arvadostest/run_servers.go
index cad1691..a175136 100644
--- a/sdk/go/arvadostest/run_servers.go
+++ b/sdk/go/arvadostest/run_servers.go
@@ -99,11 +99,18 @@ func StopAPI() {
}
func StartKeep() {
+ StartKeepAdditional(false)
+}
+
+func StartKeepAdditional(keepExisting bool) {
cwd, _ := os.Getwd()
defer os.Chdir(cwd)
chdirToPythonTests()
cmd := exec.Command("python", "run_test_server.py", "start_keep")
+ if keepExisting {
+ cmd = exec.Command("python", "run_test_server.py", "start_keep", "--keep_existing", "true")
+ }
stderr, err := cmd.StderrPipe()
if err != nil {
log.Fatalf("Setting up stderr pipe: %s", err)
diff --git a/sdk/python/tests/run_test_server.py b/sdk/python/tests/run_test_server.py
index 5d0c42a..d5d1874 100644
--- a/sdk/python/tests/run_test_server.py
+++ b/sdk/python/tests/run_test_server.py
@@ -324,7 +324,8 @@ def _start_keep(n, keep_args):
return port
def run_keep(blob_signing_key=None, enforce_permissions=False):
- stop_keep()
+ if args.keep_existing is None:
+ stop_keep()
keep_args = {}
if not blob_signing_key:
@@ -344,12 +345,16 @@ def run_keep(blob_signing_key=None, enforce_permissions=False):
host=os.environ['ARVADOS_API_HOST'],
token=os.environ['ARVADOS_API_TOKEN'],
insecure=True)
+
for d in api.keep_services().list().execute()['items']:
api.keep_services().delete(uuid=d['uuid']).execute()
for d in api.keep_disks().list().execute()['items']:
api.keep_disks().delete(uuid=d['uuid']).execute()
- for d in range(0, 2):
+ start_index = 0
+ if args.keep_existing is not None:
+ start_index = 2
+ for d in range(start_index, start_index+2):
port = _start_keep(d, keep_args)
svc = api.keep_services().create(body={'keep_service': {
'uuid': 'zzzzz-bi6l4-keepdisk{:07d}'.format(d),
@@ -374,6 +379,9 @@ def _stop_keep(n):
def stop_keep():
_stop_keep(0)
_stop_keep(1)
+ # We may have created 2 additional keep servers when keep_existing is used
+ _stop_keep(2)
+ _stop_keep(3)
def run_keep_proxy():
if 'ARVADOS_TEST_PROXY_SERVICES' in os.environ:
@@ -595,6 +603,7 @@ if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('action', type=str, help="one of {}".format(actions))
parser.add_argument('--auth', type=str, metavar='FIXTURE_NAME', help='Print authorization info for given api_client_authorizations fixture')
+ parser.add_argument('--keep_existing', type=str, help="Used to add additional keep servers, without terminating existing servers")
args = parser.parse_args()
if args.action not in actions:
diff --git a/tools/keep-rsync/.gitignore b/tools/keep-rsync/.gitignore
new file mode 100644
index 0000000..5ee7f3b
--- /dev/null
+++ b/tools/keep-rsync/.gitignore
@@ -0,0 +1 @@
+keep-rsync
diff --git a/tools/keep-rsync/keep-rsync.go b/tools/keep-rsync/keep-rsync.go
new file mode 100644
index 0000000..eff8b9c
--- /dev/null
+++ b/tools/keep-rsync/keep-rsync.go
@@ -0,0 +1,141 @@
+package main
+
+import (
+ "flag"
+ "git.curoverse.com/arvados.git/sdk/go/arvadosclient"
+ "git.curoverse.com/arvados.git/sdk/go/keepclient"
+ "io/ioutil"
+ "log"
+ "strings"
+)
+
+// keep-rsync arguments
+var (
+ srcConfig map[string]string
+ destConfig map[string]string
+ srcKeepServicesJSON string
+ destKeepServicesJSON string
+ replications int
+ prefix string
+)
+
+func main() {
+ var srcConfigFile string
+ var destConfigFile string
+
+ flag.StringVar(
+ &srcConfigFile,
+ "src-config-file",
+ "",
+ "Source configuration filename with full path that contains "+
+ "an ARVADOS_API_TOKEN which is a valid datamanager token recognized by the source keep servers, "+
+ "ARVADOS_API_HOST, ARVADOS_API_HOST_INSECURE, and ARVADOS_BLOB_SIGNING_KEY.")
+
+ flag.StringVar(
+ &destConfigFile,
+ "dest-config-file",
+ "",
+ "Destination configuration filename with full path that contains "+
+ "an ARVADOS_API_TOKEN which is a valid datamanager token recognized by the destination keep servers, "+
+ "ARVADOS_API_HOST, ARVADOS_API_HOST_INSECURE, and ARVADOS_BLOB_SIGNING_KEY.")
+
+ flag.StringVar(
+ &srcKeepServicesJSON,
+ "src-keep-services-json",
+ "",
+ "An optional list of available source keepservices. "+
+ "If not provided, this list is obtained from api server configured in src-config-file.")
+
+ flag.StringVar(
+ &destKeepServicesJSON,
+ "dest-keep-services-json",
+ "",
+ "An optional list of available destination keepservices. "+
+ "If not provided, this list is obtained from api server configured in dest-config-file.")
+
+ flag.IntVar(
+ &replications,
+ "replications",
+ 3,
+ "Number of replications to write to the destination.")
+
+ flag.StringVar(
+ &prefix,
+ "prefix",
+ "",
+ "Index prefix")
+
+ flag.Parse()
+
+ var err error
+
+ if srcConfigFile == "" {
+ log.Fatal("-src-config-file must be specified.")
+ }
+ srcConfig, err = readConfigFromFile(srcConfigFile)
+ if err != nil {
+ log.Fatal("Error reading source configuration: %s", err.Error())
+ }
+
+ if destConfigFile == "" {
+ log.Fatal("-dest-config-file must be specified.")
+ }
+ destConfig, err = readConfigFromFile(destConfigFile)
+ if err != nil {
+ log.Fatal("Error reading destination configuration: %s", err.Error())
+ }
+
+ err = initializeKeepRsync()
+ if err != nil {
+ log.Fatal("Error configurating keep-rsync: %s", err.Error())
+ }
+}
+
+// Reads config from file
+func readConfigFromFile(filename string) (map[string]string, error) {
+ content, err := ioutil.ReadFile(filename)
+ if err != nil {
+ return nil, err
+ }
+
+ config := make(map[string]string)
+ lines := strings.Split(string(content), "\n")
+ for _, line := range lines {
+ if line == "" {
+ continue
+ }
+ kv := strings.Split(line, "=")
+ config[kv[0]] = kv[1]
+ }
+ return config, nil
+}
+
+// keep-rsync source and destination clients
+var (
+ arvSrc arvadosclient.ArvadosClient
+ arvDest arvadosclient.ArvadosClient
+ kcSrc *keepclient.KeepClient
+ kcDest *keepclient.KeepClient
+)
+
+// Initializes keep-rsync using the config provided
+func initializeKeepRsync() (err error) {
+ arvSrc, err = arvadosclient.MakeArvadosClientWithConfig(srcConfig)
+ if err != nil {
+ return
+ }
+
+ arvDest, err = arvadosclient.MakeArvadosClientWithConfig(destConfig)
+ if err != nil {
+ return
+ }
+
+ kcSrc, err = keepclient.MakeKeepClient(&arvSrc)
+ if err != nil {
+ return
+ }
+
+ kcDest, err = keepclient.MakeKeepClient(&arvDest)
+
+ return
+}
diff --git a/tools/keep-rsync/keep-rsync_test.go b/tools/keep-rsync/keep-rsync_test.go
new file mode 100644
index 0000000..e2b1f0f
--- /dev/null
+++ b/tools/keep-rsync/keep-rsync_test.go
@@ -0,0 +1,117 @@
+package main
+
+import (
+ "crypto/md5"
+ "fmt"
+ "io/ioutil"
+ "os"
+ "testing"
+
+ "git.curoverse.com/arvados.git/sdk/go/arvadostest"
+ "git.curoverse.com/arvados.git/sdk/go/keepclient"
+
+ . "gopkg.in/check.v1"
+)
+
+// Gocheck boilerplate
+func Test(t *testing.T) {
+ TestingT(t)
+}
+
+// Gocheck boilerplate
+var _ = Suite(&ServerRequiredSuite{})
+
+// Tests that require the Keep server running
+type ServerRequiredSuite struct{}
+
+func (s *ServerRequiredSuite) SetUpSuite(c *C) {
+}
+
+func (s *ServerRequiredSuite) SetUpTest(c *C) {
+ arvadostest.ResetEnv()
+}
+
+func (s *ServerRequiredSuite) TearDownSuite(c *C) {
+ arvadostest.StopKeep()
+ arvadostest.StopAPI()
+}
+
+// Testing keep-rsync needs two sets of keep services: src and dest.
+// The test setup hence tweaks keep-rsync initialzation to achieve this.
+// First invoke initializeKeepRsync and then invoke StartKeepAdditional
+// to create the keep servers to be used as destination.
+func setupRsync(c *C) {
+ // srcConfig
+ srcConfig = make(map[string]string)
+ srcConfig["ARVADOS_API_HOST"] = os.Getenv("ARVADOS_API_HOST")
+ srcConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
+ srcConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
+
+ // destConfig
+ destConfig = make(map[string]string)
+ destConfig["ARVADOS_API_HOST"] = os.Getenv("ARVADOS_API_HOST")
+ destConfig["ARVADOS_API_TOKEN"] = os.Getenv("ARVADOS_API_TOKEN")
+ destConfig["ARVADOS_API_HOST_INSECURE"] = os.Getenv("ARVADOS_API_HOST_INSECURE")
+
+ arvadostest.StartAPI()
+ arvadostest.StartKeep()
+
+ // initialize keep-rsync
+ err := initializeKeepRsync()
+ c.Assert(err, Equals, nil)
+
+ // Create two more keep servers to be used as destination
+ arvadostest.StartKeepAdditional(true)
+
+ // load kcDest
+ kcDest, err = keepclient.MakeKeepClient(&arvDest)
+ c.Assert(err, Equals, nil)
+}
+
+// Test readConfigFromFile method
+func (s *ServerRequiredSuite) TestReadConfigFromFile(c *C) {
+ // Setup a test config file
+ file, err := ioutil.TempFile(os.TempDir(), "config")
+ c.Assert(err, Equals, nil)
+ defer os.Remove(file.Name())
+
+ fileContent := "ARVADOS_API_HOST=testhost\n"
+ fileContent += "ARVADOS_API_TOKEN=testtoken\n"
+ fileContent += "ARVADOS_API_HOST_INSECURE=true"
+
+ _, err = file.Write([]byte(fileContent))
+
+ // Invoke readConfigFromFile method with this test filename
+ config, err := readConfigFromFile(file.Name())
+ c.Assert(err, Equals, nil)
+ c.Assert(config["ARVADOS_API_HOST"], Equals, "testhost")
+ c.Assert(config["ARVADOS_API_TOKEN"], Equals, "testtoken")
+ c.Assert(config["ARVADOS_API_HOST_INSECURE"], Equals, "true")
+ c.Assert(config["EXTERNAL_CLIENT"], Equals, "")
+}
+
+// Test keep-rsync initialization, with src and dest keep servers.
+// Do a Put and Get in src, both of which should succeed.
+// Do a Get in dest for the same hash, which should raise block not found error.
+func (s *ServerRequiredSuite) TestRsyncPutInSrc_GetFromDestShouldFail(c *C) {
+ setupRsync(c)
+
+ // Put a block in src using kcSrc and Get it
+ data := []byte("test-data")
+ hash := fmt.Sprintf("%x", md5.Sum(data))
+
+ hash2, rep, err := kcSrc.PutB(data)
+ c.Check(hash2, Matches, fmt.Sprintf(`^%s\+9(\+.+)?$`, hash))
+ c.Check(rep, Equals, 2)
+ c.Check(err, Equals, nil)
+
+ reader, blocklen, _, err := kcSrc.Get(hash)
+ c.Assert(err, Equals, nil)
+ c.Check(blocklen, Equals, int64(9))
+ all, err := ioutil.ReadAll(reader)
+ c.Check(all, DeepEquals, data)
+
+ // Get using kcDest should fail with NotFound error
+ _, _, _, err = kcDest.Get(hash)
+ c.Assert(err.Error(), Equals, "Block not found")
+}
-----------------------------------------------------------------------
hooks/post-receive
--
More information about the arvados-commits
mailing list