func (s *CephProvider) StopTask(req models.RpcRequest, resp *models.RpcResponse) error {
ctxt := req.RpcRequestContext
task_id_str := req.RpcRequestVars["task-id"]
logger.Get().Debug(fmt.Sprintf("%s-Stopping sub-task: %s", ctxt, task_id_str))
task_id, err := uuid.Parse(task_id_str)
if err != nil {
logger.Get().Error("%s-Error parsing the task id: %s. error: %v", ctxt, task_id_str, err)
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Error parsing the task id: %s", task_id_str))
return err
}
// Stop the task
if ok, err := task.GetTaskManager().Stop(*task_id); !ok || err != nil {
logger.Get().Error("%s-Failed to stop the task: %v. error: %v", ctxt, task_id, err)
*resp = utils.WriteResponse(http.StatusInternalServerError, "Failed to stop task")
return err
} else {
*resp = utils.WriteResponse(http.StatusOK, "Done")
}
return nil
}
func (s *CephProvider) UpdateStorage(req models.RpcRequest, resp *models.RpcResponse) error {
ctxt := req.RpcRequestContext
cluster_id_str := req.RpcRequestVars["cluster-id"]
cluster_id, err := uuid.Parse(cluster_id_str)
if err != nil {
logger.Get().Error(
"%s - Error parsing the cluster id: %s. error: %v",
ctxt,
cluster_id_str,
err)
*resp = utils.WriteResponse(
http.StatusBadRequest,
fmt.Sprintf(
"Error parsing the cluster id: %s",
cluster_id_str))
return err
}
storage_id_str := req.RpcRequestVars["storage-id"]
storage_id, err := uuid.Parse(storage_id_str)
if err != nil {
logger.Get().Error(
"%s - Error parsing the storage id: %s. error: %v",
ctxt,
storage_id_str,
err)
*resp = utils.WriteResponse(
http.StatusBadRequest,
fmt.Sprintf(
"Error parsing the storage id: %s",
storage_id_str))
return err
}
var request models.AddStorageRequest
if err := json.Unmarshal(req.RpcRequestData, &request); err != nil {
logger.Get().Error(
"%s - Unbale to parse the request. error: %v",
ctxt,
err)
*resp = utils.WriteResponse(
http.StatusBadRequest,
fmt.Sprintf(
"Unbale to parse the request. error: %v",
err))
return err
}
sessionCopy := db.GetDatastore().Copy()
defer sessionCopy.Close()
coll := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_STORAGE)
var storage models.Storage
if err := coll.Find(bson.M{"storageid": *storage_id}).One(&storage); err != nil {
logger.Get().Error(
"%s - Error getting detals of storage: %v on cluster: %v. error: %v",
ctxt,
*storage_id,
*cluster_id,
err)
*resp = utils.WriteResponse(
http.StatusInternalServerError,
fmt.Sprintf(
"Error getting the details of storage: %v",
*storage_id))
return err
}
id, err := strconv.Atoi(storage.Options["id"])
if err != nil {
logger.Get().Error(
"%s - Error getting id of the pool: %v of cluster: %v. error: %v",
ctxt,
*storage_id,
*cluster_id,
err)
*resp = utils.WriteResponse(
http.StatusInternalServerError,
fmt.Sprintf(
"Error getting id of the pool: %v",
*storage_id))
return err
}
asyncTask := func(t *task.Task) {
sessionCopy := db.GetDatastore().Copy()
defer sessionCopy.Close()
coll := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_STORAGE)
for {
select {
case <-t.StopCh:
return
default:
t.UpdateStatus("Started ceph provider pool updation: %v", t.ID)
if request.Name != "" && (request.Replicas != 0 || len(request.Options) != 0) {
utils.FailTask(
fmt.Sprintf(
"Invalid mix of fields to update for storage: %v of cluster: %v. "+
"Name change cannot be mixed with other changes.",
*storage_id,
*cluster_id),
fmt.Errorf("%s-Invalid mix of fields to update", ctxt),
t)
return
}
for key := range request.Options {
if ok := skyring_util.StringInSlice(key, validConfigs); !ok {
utils.FailTask(
fmt.Sprintf(
"Invalid configuration: %s mentioned for storage: %v of cluster: %v",
key,
*storage_id,
*cluster_id),
fmt.Errorf("%s-%v", ctxt, err),
t)
return
}
}
t.UpdateStatus("Getting a radom mon from cluster")
monnode, err := GetCalamariMonNode(*cluster_id, ctxt)
if err != nil {
utils.FailTask(
fmt.Sprintf(
"Error getting mon node from cluster: %v",
*cluster_id),
fmt.Errorf("%s-%v", ctxt, err),
t)
return
}
var updatedFields = make(map[string]interface{})
if request.Name != "" {
updatedFields["name"] = request.Name
}
if request.Replicas != 0 {
updatedFields["size"] = request.Replicas
}
if request.QuotaEnabled {
for key, value := range request.QuotaParams {
reqVal, _ := strconv.ParseUint(value, 10, 64)
updatedFields[key] = uint64(reqVal)
}
} else {
if request.QuotaParams["quota_max_objects"] == "0" && request.QuotaParams["quota_max_bytes"] == "0" {
updatedFields["quota_max_objects"] = 0
updatedFields["quota_max_bytes"] = 0
}
}
for key, value := range request.Options {
reqVal, _ := strconv.ParseUint(value, 10, 32)
updatedFields[key] = uint(reqVal)
}
t.UpdateStatus("Updating pool details")
ok, err := cephapi_backend.UpdatePool(
monnode.Hostname,
*cluster_id, id,
updatedFields,
ctxt)
if err != nil || !ok {
utils.FailTask(
fmt.Sprintf(
"Error setting the configurations for storage: %v on cluster: %v",
*storage_id,
*cluster_id),
fmt.Errorf("%s-%v", ctxt, err),
t)
return
}
var filter bson.M = make(map[string]interface{})
var updates bson.M = make(map[string]interface{})
filter["storageid"] = *storage_id
filter["clusterid"] = *cluster_id
if request.Name != "" {
updates["name"] = request.Name
}
if request.Replicas != 0 {
updates["replicas"] = request.Replicas
}
if request.QuotaEnabled {
updates["quotaenabled"] = true
params := make(map[string]string)
for key, value := range request.QuotaParams {
params[key] = string(value)
}
updates["quotaparams"] = params
} else {
if request.QuotaParams["quota_max_objects"] == "0" && request.QuotaParams["quota_max_bytes"] == "0" {
updates["quotaenabled"] = false
updates["quotaparams"] = map[string]string{}
}
}
if value, ok := request.Options["pg_num"]; ok {
updates["options.pgp_num"] = value
}
t.UpdateStatus("Persisting pool updates in DB")
if err := coll.Update(filter, bson.M{"$set": updates}); err != nil {
utils.FailTask(
fmt.Sprintf(
"Error updating storage entity: %v of cluster: %v",
*storage_id,
*cluster_id),
fmt.Errorf("%s-%v", ctxt, err),
t)
}
cluster, err := getCluster(*cluster_id)
if err != nil {
logger.Get().Error("%s - Failed to get details of cluster: %s. error: %v", ctxt, *cluster_id, err)
} else {
initMonitoringRoutines(ctxt, cluster, (*monnode).Hostname, []interface{}{FetchOSDStats})
UpdatePgNumToSummaries(cluster, ctxt)
}
if _, ok := updates["name"]; ok {
coll := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_BLOCK_DEVICES)
if _, err := coll.UpdateAll(filter, bson.M{"$set": bson.M{"storagename": updates["name"]}}); err != nil && err != mgo.ErrNotFound {
utils.FailTask(
fmt.Sprintf(
"Storage name has changed for storage:%v. Error while updating this info for RBDs in cluster:%v",
*storage_id,
*cluster_id),
fmt.Errorf("%s-%v", ctxt, err),
t)
}
}
t.UpdateStatus("Success")
t.Done(models.TASK_STATUS_SUCCESS)
return
}
}
}
if taskId, err := bigfin_task.GetTaskManager().Run(
bigfin_conf.ProviderName,
"CEPH-UpdateStorage",
asyncTask,
nil,
nil,
nil); err != nil {
logger.Get().Error(
"%s-Task creation failed for update storage on cluster: %v. error: %v",
ctxt,
*cluster_id,
err)
*resp = utils.WriteResponse(
http.StatusInternalServerError,
"Task creation failed for storage update")
return err
} else {
*resp = utils.WriteAsyncResponse(
taskId,
fmt.Sprintf(
"Task Created for update storage on cluster: %v",
*cluster_id),
[]byte{})
}
return nil
}
func (s *CephProvider) CreateStorage(req models.RpcRequest, resp *models.RpcResponse) error {
ctxt := req.RpcRequestContext
var request models.AddStorageRequest
if err := json.Unmarshal(req.RpcRequestData, &request); err != nil {
logger.Get().Error("%s - Unbale to parse the request. error: %v", ctxt, err)
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Unbale to parse the request. error: %v", err))
return err
}
// Create the storage pool
cluster_id_str := req.RpcRequestVars["cluster-id"]
cluster_id, err := uuid.Parse(cluster_id_str)
if err != nil {
logger.Get().Error("%s - Error parsing the cluster id: %s. error: %v", ctxt, cluster_id_str, err)
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Error parsing the cluster id: %s", cluster_id_str))
return err
}
asyncTask := func(t *task.Task) {
sessionCopy := db.GetDatastore().Copy()
defer sessionCopy.Close()
for {
select {
case <-t.StopCh:
return
default:
var cluster models.Cluster
t.UpdateStatus("Started ceph provider storage creation: %v", t.ID)
t.UpdateStatus("Getting cluster details")
// Get cluster details
coll := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_STORAGE_CLUSTERS)
if err := coll.Find(bson.M{"clusterid": *cluster_id}).One(&cluster); err != nil {
utils.FailTask(fmt.Sprintf("Error getting the cluster details for :%v", *cluster_id), fmt.Errorf("%s - %v", ctxt, err), t)
return
}
t.UpdateStatus("Getting a mon from cluster")
monnode, err := GetCalamariMonNode(*cluster_id, ctxt)
if err != nil {
utils.FailTask(fmt.Sprintf("Error getting mon node details for cluster: %v", *cluster_id), fmt.Errorf("%s - %v", ctxt, err), t)
return
}
poolId, ok := createPool(ctxt, *cluster_id, request, t)
if !ok {
return
}
if request.Type != models.STORAGE_TYPE_ERASURE_CODED && len(request.BlockDevices) > 0 {
createBlockDevices(ctxt, monnode.Hostname, cluster, *poolId, request, t)
}
t.UpdateStatus("Syncing SLUs")
if err := SyncOsdStatus(*cluster_id, ctxt); err != nil {
utils.FailTask("Error syncing SLUs", err, t)
return
}
initMonitoringRoutines(ctxt, cluster, (*monnode).Hostname, []interface{}{FetchObjectCount})
_, cStats, err := updateClusterStats(ctxt, cluster, (*monnode).Hostname)
if err == nil {
updateStatsToPools(ctxt, cStats, cluster.ClusterId)
}
skyring_util.UpdateStorageCountToSummaries(ctxt, cluster)
UpdateObjectCountToSummaries(ctxt, cluster)
t.UpdateStatus("Success")
t.Done(models.TASK_STATUS_SUCCESS)
return
}
}
}
if taskId, err := bigfin_task.GetTaskManager().Run(
bigfin_conf.ProviderName,
"CEPH-CreateStorage",
asyncTask,
nil,
nil,
nil); err != nil {
logger.Get().Error("%s-Task creation failed for create storage %s on cluster: %v. error: %v", ctxt, request.Name, *cluster_id, err)
*resp = utils.WriteResponse(http.StatusInternalServerError, "Task creation failed for storage creation")
return err
} else {
*resp = utils.WriteAsyncResponse(taskId, fmt.Sprintf("Task Created for create storage %s on cluster: %v", request.Name, *cluster_id), []byte{})
}
return nil
}
func (s *CephProvider) RemoveStorage(req models.RpcRequest, resp *models.RpcResponse) error {
ctxt := req.RpcRequestContext
cluster_id_str := req.RpcRequestVars["cluster-id"]
cluster_id, err := uuid.Parse(cluster_id_str)
if err != nil {
logger.Get().Error("%s - Error parsing the cluster id: %s. error: %v", ctxt, cluster_id_str, err)
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Error parsing the cluster id: %s", cluster_id_str))
return err
}
storage_id_str := req.RpcRequestVars["storage-id"]
storage_id, err := uuid.Parse(storage_id_str)
if err != nil {
logger.Get().Error("%s - Error parsing the storage id: %s. error: %v", ctxt, storage_id_str, err)
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Error parsing the storage id: %s", storage_id_str))
return err
}
asyncTask := func(t *task.Task) {
sessionCopy := db.GetDatastore().Copy()
defer sessionCopy.Close()
for {
select {
case <-t.StopCh:
return
default:
t.UpdateStatus("Started ceph provider pool deletion: %v", t.ID)
// Get the storage details
var storage models.Storage
var cluster models.Cluster
t.UpdateStatus("Getting details of cluster")
coll := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_STORAGE_CLUSTERS)
if err := coll.Find(bson.M{"clusterid": *cluster_id}).One(&cluster); err != nil {
utils.FailTask("Error getting details of cluster", fmt.Errorf("%s - %v", ctxt, err), t)
return
}
t.UpdateStatus("Getting details of storage")
coll1 := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_STORAGE)
if err := coll1.Find(bson.M{"clusterid": *cluster_id, "storageid": *storage_id}).One(&storage); err != nil {
utils.FailTask("Error getting details of storage", fmt.Errorf("%s - %v", ctxt, err), t)
return
}
t.UpdateStatus("Getting a mon from cluster")
monnode, err := GetCalamariMonNode(*cluster_id, ctxt)
if err != nil {
utils.FailTask("Error getting a mon node for cluster", fmt.Errorf("%s - %v", ctxt, err), t)
return
}
poolId, err := strconv.Atoi(storage.Options["id"])
if err != nil {
utils.FailTask("Error getting id of storage", fmt.Errorf("%s - %v", ctxt, err), t)
return
}
t.UpdateStatus("Deleting storage")
ok, err := cephapi_backend.RemovePool(monnode.Hostname, *cluster_id, cluster.Name, storage.Name, poolId, ctxt)
if err != nil || !ok {
utils.FailTask(fmt.Sprintf("Deletion of storage %v failed on cluster: %s", *storage_id, *cluster_id), fmt.Errorf("%s - %v", ctxt, err), t)
return
} else {
t.UpdateStatus("Removing the block devices (if any) for storage entoty")
coll2 := sessionCopy.DB(conf.SystemConfig.DBConfig.Database).C(models.COLL_NAME_BLOCK_DEVICES)
if _, err := coll2.RemoveAll(bson.M{"clusterid": *cluster_id, "storageid": *storage_id}); err != nil {
utils.FailTask(fmt.Sprintf("Error removing block devices for storage %v from DB for cluster: %d", *storage_id, *cluster_id), fmt.Errorf("%s - %v", ctxt, err), t)
return
}
t.UpdateStatus("Removing the storage entity from DB")
if err := coll1.Remove(bson.M{"clusterid": *cluster_id, "storageid": *storage_id}); err != nil {
utils.FailTask(fmt.Sprintf("Error removing storage entity from DB for cluster: %d", *cluster_id), fmt.Errorf("%s - %v", ctxt, err), t)
return
}
}
t.UpdateStatus("Syncing SLUs")
if err := SyncOsdStatus(*cluster_id, ctxt); err != nil {
utils.FailTask("Error syncing SLUs", err, t)
return
}
skyring_util.UpdateStorageCountToSummaries(ctxt, cluster)
UpdateObjectCountToSummaries(ctxt, cluster)
t.UpdateStatus("Success")
t.Done(models.TASK_STATUS_SUCCESS)
return
}
}
}
if taskId, err := bigfin_task.GetTaskManager().Run(
bigfin_conf.ProviderName,
"CEPH-DeleteStorage",
asyncTask,
nil,
nil,
nil); err != nil {
logger.Get().Error("%s-Task creation failed for delete storage on cluster: %v. error: %v", ctxt, *cluster_id, err)
*resp = utils.WriteResponse(http.StatusInternalServerError, "Task creation failed for storage deletion")
return err
} else {
*resp = utils.WriteAsyncResponse(taskId, fmt.Sprintf("Task Created for delete storage on cluster: %v", *cluster_id), []byte{})
}
return nil
}
func (s *CephProvider) ImportCluster(req models.RpcRequest, resp *models.RpcResponse) error {
var request models.ImportClusterRequest
ctxt := req.RpcRequestContext
if err := json.Unmarshal(req.RpcRequestData, &request); err != nil {
logger.Get().Error(
fmt.Sprintf("%s-Unbale to parse the import cluster request. error: %v",
ctxt,
err))
*resp = utils.WriteResponse(http.StatusBadRequest, fmt.Sprintf("Unbale to parse the request. error: %v", err))
return err
}
asyncTask := func(t *task.Task) {
// Get cluster details and populate
t.UpdateStatus("Updating cluster generic dtails (name/id)")
cluster_uuid, clusterName, err := PopulateClusterDetails(request.BootstrapNode, ctxt)
if err != nil {
utils.FailTask(
"Failed to fetch and populate cluster details",
err,
t)
return
}
setClusterState(*cluster_uuid, models.CLUSTER_STATE_CREATING, ctxt)
// Get the cluster network details and populate
t.UpdateStatus("Updating cluster network details")
if err := PopulateClusterNetworkDetails(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
utils.FailTask(
fmt.Sprintf(
"Error fetching and populating network details for cluster: %s",
clusterName),
err,
t)
setClusterState(*cluster_uuid, models.CLUSTER_STATE_FAILED, ctxt)
return
}
// Get and populate cluster status details
t.UpdateStatus("Updating cluster status")
if err := PopulateClusterStatus(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
utils.FailTask(
fmt.Sprintf(
"Failed to fetch and populate status details for cluster: %s",
clusterName),
err,
t)
setClusterState(*cluster_uuid, models.CLUSTER_STATE_FAILED, ctxt)
return
}
// Get and update nodes of the cluster
t.UpdateStatus("Updating cluster nodes")
failedNodes, err := PopulateClusterNodes(request.BootstrapNode, *cluster_uuid, request.Nodes, ctxt)
if err != nil {
utils.FailTask(
fmt.Sprintf(
"Failed populating node details for cluster: %s",
clusterName),
err,
t)
setClusterState(*cluster_uuid, models.CLUSTER_STATE_FAILED, ctxt)
return
}
if len(failedNodes) > 0 {
t.UpdateStatus("Failed to updated details of nodes: %v", failedNodes)
}
// Get and update storage pools
t.UpdateStatus("Updating cluster storages")
if err := PopulateStoragePools(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
t.UpdateStatus("Failed populating storage pools details")
logger.Get().Warning(
"%s-Failed populating storage pools details for cluster: %s. error: %v",
ctxt,
clusterName,
err)
}
// Get and update OSDs
t.UpdateStatus("Updating cluster SLUs")
if err := PopulateClusterOSDs(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
t.UpdateStatus("Failed populating OSD details")
logger.Get().Warning(
"%s-Failed populating OSD details for cluster: %s. error: %v",
ctxt,
clusterName,
err)
}
// Check and disable auto expand if colocated journals
checkAndUpdateAutoExpand(t, *cluster_uuid, ctxt)
// Get and update block devices
t.UpdateStatus("Updating cluster block devices")
if err := PopulateBlockDevices(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
t.UpdateStatus("Failed populating block devices details")
logger.Get().Warning(
"%s-Failed populating block devices details for cluster: %s. error: %v",
ctxt,
clusterName,
err)
}
setClusterState(*cluster_uuid, models.CLUSTER_STATE_ACTIVE, ctxt)
// Create default EC profiles
t.UpdateStatus("Creating default EC profiles")
if ok, err := CreateDefaultECProfiles(
ctxt,
request.BootstrapNode,
*cluster_uuid); !ok || err != nil {
logger.Get().Error(
"%s-Error creating default EC profiles for cluster: %s. error: %v",
ctxt,
clusterName,
err)
t.UpdateStatus("Could not create default EC profile")
}
//Populate the CRUSH Details
t.UpdateStatus("Setting up the CRUSH Map")
if err := poplateCrushDetails(request.BootstrapNode, *cluster_uuid, ctxt); err != nil {
t.UpdateStatus("Failed to update Crush map")
}
// Update the notification details
t.UpdateStatus("Creating notification subscription configuartions")
var bigfin_notifications []models.NotificationSubscription
for _, notification := range models.Notifications {
bigfin_notifications = append(bigfin_notifications, notification)
}
for _, notification := range bigfin_models.Notifications {
bigfin_notifications = append(bigfin_notifications, notification)
}
sessionCopy := db.GetDatastore().Copy()
defer sessionCopy.Close()
notifSubsColl := sessionCopy.
DB(conf.SystemConfig.DBConfig.Database).
C(models.COLL_NAME_CLUSTER_NOTIFICATION_SUBSCRIPTIONS)
if err := notifSubsColl.Insert(
models.ClusterNotificationSubscription{
ClusterId: *cluster_uuid,
Notifications: bigfin_notifications}); err != nil {
logger.Get().Error(
"%s-Error persisting the default notification subscriptions on cluster %s. error %v",
ctxt,
clusterName,
err)
t.UpdateStatus("Could not create notification subscription configuartions")
}
t.UpdateStatus("Success")
t.Done(models.TASK_STATUS_SUCCESS)
return
}
if taskId, err := bigfin_task.GetTaskManager().Run(
bigfin_conf.ProviderName,
fmt.Sprintf("%s-Import Cluster", bigfin_conf.ProviderName),
asyncTask,
nil,
nil,
nil); err != nil {
logger.Get().Error(
"%s - Task creation failed for import cluster. error: %v",
ctxt,
err)
*resp = utils.WriteResponse(
http.StatusInternalServerError,
"Task creation failed for import cluster")
return err
} else {
*resp = utils.WriteAsyncResponse(
taskId,
"Task Created for import cluster",
[]byte{})
}
return nil
}