Add large bucket support for erasure coded backend (#5160)

This PR implements an object layer which
combines input erasure sets of XL layers
into a unified namespace.

This object layer extends the existing
erasure coded implementation, it is assumed
in this design that providing > 16 disks is
a static configuration as well i.e if you started
the setup with 32 disks with 4 sets 8 disks per
pack then you would need to provide 4 sets always.

Some design details and restrictions:

- Objects are distributed using consistent ordering
  to a unique erasure coded layer.
- Each pack has its own dsync so locks are synchronized
  properly at pack (erasure layer).
- Each pack still has a maximum of 16 disks
  requirement, you can start with multiple
  such sets statically.
- Static sets set of disks and cannot be
  changed, there is no elastic expansion allowed.
- Static sets set of disks and cannot be
  changed, there is no elastic removal allowed.
- ListObjects() across sets can be noticeably
  slower since List happens on all servers,
  and is merged at this sets layer.

Fixes #5465
Fixes #5464
Fixes #5461
Fixes #5460
Fixes #5459
Fixes #5458
Fixes #5460
Fixes #5488
Fixes #5489
Fixes #5497
Fixes #5496
This commit is contained in:
Harshavardhana
2018-02-15 17:45:57 -08:00
committed by kannappanr
parent dd80256151
commit fb96779a8a
82 changed files with 5046 additions and 4771 deletions
+31 -31
View File
@@ -46,7 +46,16 @@ var serverCmd = cli.Command{
{{.HelpName}} - {{.Usage}}
USAGE:
{{.HelpName}} {{if .VisibleFlags}}[FLAGS] {{end}}PATH [PATH...]
{{.HelpName}} {{if .VisibleFlags}}[FLAGS] {{end}}DIR1 [DIR2..]
{{.HelpName}} {{if .VisibleFlags}}[FLAGS] {{end}}DIR{1...64}
DIR:
DIR points to a directory on a filesystem. When you want to combine
multiple drives into a single large system, pass one directory per
filesystem separated by space. You may also use a '...' convention
to abbreviate the directory arguments. Remote directories in a
distributed setup are encoded as HTTP(s) URIs.
{{if .VisibleFlags}}
FLAGS:
{{range .VisibleFlags}}{{.}}
@@ -72,16 +81,24 @@ EXAMPLES:
2. Start minio server bound to a specific ADDRESS:PORT.
$ {{.HelpName}} --address 192.168.1.101:9000 /home/shared
3. Start erasure coded minio server on a 12 disks server.
3. Start minio server on a 12 disks server.
$ {{.HelpName}} /mnt/export1/ /mnt/export2/ /mnt/export3/ /mnt/export4/ \
/mnt/export5/ /mnt/export6/ /mnt/export7/ /mnt/export8/ /mnt/export9/ \
/mnt/export10/ /mnt/export11/ /mnt/export12/
4. Start erasure coded distributed minio server on a 4 node setup with 1 drive each. Run following commands on all the 4 nodes.
4. Start distributed minio server on a 4 node setup with 1 drive each. Run following commands on all the 4 nodes.
$ export MINIO_ACCESS_KEY=minio
$ export MINIO_SECRET_KEY=miniostorage
$ {{.HelpName}} http://192.168.1.11/mnt/export/ http://192.168.1.12/mnt/export/ \
http://192.168.1.13/mnt/export/ http://192.168.1.14/mnt/export/
5. Start minio server on 64 disks server.
$ {{.HelpName}} /mnt/export{1...64}
6. Start distributed minio server on an 8 node setup with 8 drives each. Run following command on all the 8 nodes.
$ export MINIO_ACCESS_KEY=minio
$ export MINIO_SECRET_KEY=miniostorage
$ {{.HelpName}} http://node{1...8}.example.com/mnt/export/{1...8}
`,
}
@@ -96,8 +113,13 @@ func serverHandleCmdArgs(ctx *cli.Context) {
var setupType SetupType
var err error
globalMinioAddr, globalEndpoints, setupType, err = CreateEndpoints(serverAddr, ctx.Args()...)
if len(ctx.Args()) > serverCommandLineArgsMax {
fatalIf(errInvalidArgument, "Invalid total number of arguments (%d) passed, supported upto 32 unique arguments", len(ctx.Args()))
}
globalMinioAddr, globalEndpoints, setupType, globalXLSetCount, globalXLSetDriveCount, err = createServerEndpoints(serverAddr, ctx.Args()...)
fatalIf(err, "Invalid command line arguments server=%s, args=%s", serverAddr, ctx.Args())
globalMinioHost, globalMinioPort = mustSplitHostPort(globalMinioAddr)
if runtime.GOOS == "darwin" {
// On macOS, if a process already listens on LOCALIPADDR:PORT, net.Listen() falls back
@@ -128,7 +150,7 @@ func serverHandleEnvVars() {
// serverMain handler called for 'minio server' command.
func serverMain(ctx *cli.Context) {
if !ctx.Args().Present() || ctx.Args().First() == "help" {
if (!ctx.IsSet("sets") && !ctx.Args().Present()) || ctx.Args().First() == "help" {
cli.ShowCommandHelpAndExit(ctx, "server", 1)
}
@@ -187,7 +209,7 @@ func serverMain(ctx *cli.Context) {
// Set nodes for dsync for distributed setup.
if globalIsDistXL {
globalDsync, err = dsync.New(newDsyncNodes(globalEndpoints))
fatalIf(err, "Unable to initialize distributed locking clients")
fatalIf(err, "Unable to initialize distributed locking on %s", globalEndpoints)
}
// Initialize name space lock.
@@ -244,39 +266,17 @@ func serverMain(ctx *cli.Context) {
// Initialize object layer with the supplied disks, objectLayer is nil upon any error.
func newObjectLayer(endpoints EndpointList) (newObject ObjectLayer, err error) {
// For FS only, directly use the disk.
isFS := len(endpoints) == 1
if isFS {
// Initialize new FS object layer.
return newFSObjectLayer(endpoints[0].Path)
}
// Initialize storage disks.
storageDisks, err := initStorageDisks(endpoints)
format, err := waitForFormatXL(endpoints[0].IsLocal, endpoints, globalXLSetCount, globalXLSetDriveCount)
if err != nil {
return nil, err
}
// Wait for formatting disks for XL backend.
var formattedDisks []StorageAPI
// First disk argument check if it is local.
firstDisk := endpoints[0].IsLocal
formattedDisks, err = waitForFormatXLDisks(firstDisk, endpoints, storageDisks)
if err != nil {
return nil, err
}
// Cleanup objects that weren't successfully written into the namespace.
if err = houseKeeping(storageDisks); err != nil {
return nil, err
}
// Once XL formatted, initialize object layer.
newObject, err = newXLObjectLayer(formattedDisks)
if err != nil {
return nil, err
}
// XL initialized, return.
return newObject, nil
return newXLSets(endpoints, format, len(format.XL.Sets), len(format.XL.Sets[0]))
}