minio/cmd/admin-handlers-pools.go

363 lines
10 KiB
Go

// Copyright (c) 2015-2021 MinIO, Inc.
//
// This file is part of MinIO Object Storage stack
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package cmd
import (
"encoding/json"
"errors"
"fmt"
"net/http"
"strings"
"github.com/minio/minio/internal/logger"
"github.com/minio/mux"
iampolicy "github.com/minio/pkg/iam/policy"
)
var (
errRebalanceDecommissionAlreadyRunning = errors.New("Rebalance cannot be started, decommission is already in progress")
errDecommissionRebalanceAlreadyRunning = errors.New("Decommission cannot be started, rebalance is already in progress")
)
func (a adminAPIHandlers) StartDecommission(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.DecommissionAdminAction)
if objectAPI == nil {
return
}
// Legacy args style such as non-ellipses style is not supported with this API.
if globalEndpoints.Legacy() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
z, ok := objectAPI.(*erasureServerPools)
if !ok || len(z.serverPools) == 1 {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
if z.IsDecommissionRunning() {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errDecommissionAlreadyRunning), r.URL)
return
}
if z.IsRebalanceStarted() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminRebalanceAlreadyStarted), r.URL)
return
}
vars := mux.Vars(r)
v := vars["pool"]
pools := strings.Split(v, ",")
poolIndices := make([]int, 0, len(pools))
for _, pool := range pools {
idx := globalEndpoints.GetPoolIdx(pool)
if idx == -1 {
// We didn't find any matching pools, invalid input
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
return
}
var pool *erasureSets
for pidx := range z.serverPools {
if pidx == idx {
pool = z.serverPools[idx]
break
}
}
if pool == nil {
// We didn't find any matching pools, invalid input
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
return
}
poolIndices = append(poolIndices, idx)
}
if len(poolIndices) > 0 && !globalEndpoints[poolIndices[0]].Endpoints[0].IsLocal {
ep := globalEndpoints[poolIndices[0]].Endpoints[0]
for nodeIdx, proxyEp := range globalProxyEndpoints {
if proxyEp.Endpoint.Host == ep.Host {
if proxyRequestByNodeIndex(ctx, w, r, nodeIdx) {
return
}
}
}
}
if err := z.Decommission(r.Context(), poolIndices...); err != nil {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
return
}
}
func (a adminAPIHandlers) CancelDecommission(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.DecommissionAdminAction)
if objectAPI == nil {
return
}
// Legacy args style such as non-ellipses style is not supported with this API.
if globalEndpoints.Legacy() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
vars := mux.Vars(r)
v := vars["pool"]
idx := globalEndpoints.GetPoolIdx(v)
if idx == -1 {
// We didn't find any matching pools, invalid input
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errInvalidArgument), r.URL)
return
}
if ep := globalEndpoints[idx].Endpoints[0]; !ep.IsLocal {
for nodeIdx, proxyEp := range globalProxyEndpoints {
if proxyEp.Endpoint.Host == ep.Host {
if proxyRequestByNodeIndex(ctx, w, r, nodeIdx) {
return
}
}
}
}
if err := pools.DecommissionCancel(ctx, idx); err != nil {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
return
}
}
func (a adminAPIHandlers) StatusPool(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.ServerInfoAdminAction, iampolicy.DecommissionAdminAction)
if objectAPI == nil {
return
}
// Legacy args style such as non-ellipses style is not supported with this API.
if globalEndpoints.Legacy() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
vars := mux.Vars(r)
v := vars["pool"]
idx := globalEndpoints.GetPoolIdx(v)
if idx == -1 {
apiErr := toAdminAPIErr(ctx, errInvalidArgument)
apiErr.Description = fmt.Sprintf("specified pool '%s' not found, please specify a valid pool", v)
// We didn't find any matching pools, invalid input
writeErrorResponseJSON(ctx, w, apiErr, r.URL)
return
}
status, err := pools.Status(r.Context(), idx)
if err != nil {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
return
}
logger.LogIf(r.Context(), json.NewEncoder(w).Encode(&status))
}
func (a adminAPIHandlers) ListPools(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.ServerInfoAdminAction, iampolicy.DecommissionAdminAction)
if objectAPI == nil {
return
}
// Legacy args style such as non-ellipses style is not supported with this API.
if globalEndpoints.Legacy() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
poolsStatus := make([]PoolStatus, len(globalEndpoints))
for idx := range globalEndpoints {
status, err := pools.Status(r.Context(), idx)
if err != nil {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
return
}
poolsStatus[idx] = status
}
logger.LogIf(r.Context(), json.NewEncoder(w).Encode(poolsStatus))
}
func (a adminAPIHandlers) RebalanceStart(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.RebalanceAdminAction)
if objectAPI == nil {
return
}
// NB rebalance-start admin API is always coordinated from first pool's
// first node. The following is required to serialize (the effects of)
// concurrent rebalance-start commands.
if ep := globalEndpoints[0].Endpoints[0]; !ep.IsLocal {
for nodeIdx, proxyEp := range globalProxyEndpoints {
if proxyEp.Endpoint.Host == ep.Host {
if proxyRequestByNodeIndex(ctx, w, r, nodeIdx) {
return
}
}
}
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok || len(pools.serverPools) == 1 {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
if pools.IsDecommissionRunning() {
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, errRebalanceDecommissionAlreadyRunning), r.URL)
return
}
if pools.IsRebalanceStarted() {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminRebalanceAlreadyStarted), r.URL)
return
}
bucketInfos, err := objectAPI.ListBuckets(ctx, BucketOptions{})
if err != nil {
writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
return
}
buckets := make([]string, 0, len(bucketInfos))
for _, bInfo := range bucketInfos {
buckets = append(buckets, bInfo.Name)
}
var id string
if id, err = pools.initRebalanceMeta(ctx, buckets); err != nil {
writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
return
}
// Rebalance routine is run on the first node of any pool participating in rebalance.
pools.StartRebalance()
b, err := json.Marshal(struct {
ID string `json:"id"`
}{ID: id})
if err != nil {
writeErrorResponseJSON(ctx, w, toAPIError(ctx, err), r.URL)
return
}
writeSuccessResponseJSON(w, b)
// Notify peers to load rebalance.bin and start rebalance routine if they happen to be
// participating pool's leader node
globalNotificationSys.LoadRebalanceMeta(ctx, true)
}
func (a adminAPIHandlers) RebalanceStatus(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.RebalanceAdminAction)
if objectAPI == nil {
return
}
// Proxy rebalance-status to first pool first node, so that users see a
// consistent view of rebalance progress even though different rebalancing
// pools may temporarily have out of date info on the others.
if ep := globalEndpoints[0].Endpoints[0]; !ep.IsLocal {
for nodeIdx, proxyEp := range globalProxyEndpoints {
if proxyEp.Endpoint.Host == ep.Host {
if proxyRequestByNodeIndex(ctx, w, r, nodeIdx) {
return
}
}
}
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
rs, err := rebalanceStatus(ctx, pools)
if err != nil {
if errors.Is(err, errRebalanceNotStarted) || errors.Is(err, errConfigNotFound) {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminRebalanceNotStarted), r.URL)
return
}
logger.LogIf(ctx, fmt.Errorf("failed to fetch rebalance status: %w", err))
writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
return
}
logger.LogIf(r.Context(), json.NewEncoder(w).Encode(rs))
}
func (a adminAPIHandlers) RebalanceStop(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
objectAPI, _ := validateAdminReq(ctx, w, r, iampolicy.RebalanceAdminAction)
if objectAPI == nil {
return
}
pools, ok := objectAPI.(*erasureServerPools)
if !ok {
writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrNotImplemented), r.URL)
return
}
// Cancel any ongoing rebalance operation
globalNotificationSys.StopRebalance(r.Context())
writeSuccessResponseHeadersOnly(w)
logger.LogIf(ctx, pools.saveRebalanceStats(GlobalContext, 0, rebalSaveStoppedAt))
}