feat: add support s3 protocol

Author: Denise Perez

README.md

@@ -220,6 +220,30 @@ nfs_setup_protocol                         = false
 nfs_setup_protocol = true
 ```
 
+## S3 Protocol Gateways
+We support creating S3 protocol gateways that will be mounted automatically to the cluster.
+<br>In order to create you need to provide the number of protocol gateways instances you want (by default the number is 0),
+
+*The amount of S3 protocol gateways should be at least 3.*
+</br>
+for example:
+```hcl
+s3_protocol_gateways_number = 3
+```
+This will automatically create 3 instances.
+<br>In addition you can supply these optional variables:
+```hcl
+s3_protocol_gateway_instance_type         = "Standard_D8_v5"
+s3_protocol_gateway_nics_num              = 2
+s3_protocol_gateway_disk_size             = 48
+s3_protocol_gateway_frontend_cores_num    = 1
+```
+
+<br>In order to create stateless clients, need to set variable:
+```hcl
+s3_setup_protocol = true
+```
+
 ## SMB Protocol Gateways
 We support creating SMB protocol gateways that will be mounted automatically to the cluster.
 <br>In order to create you need to provide the number of protocol gateways instances you want (by default the number is 0),

modules/protocol_gateways/main.tf

@@ -147,13 +147,25 @@ locals {
     key_vault_url                = var.key_vault_url
   })
 
-  protocol_script = var.protocol == "NFS" ? local.setup_nfs_protocol_script : local.setup_smb_protocol_script
+  setup_s3_protocol_script = file("${path.module}/setup_s3.sh")
+
+  setup_validation_script = templatefile("${path.module}/setup_validation.sh", {
+    gateways_number             = var.gateways_number
+    gateways_name               = var.gateways_name
+    report_function_url         = format("https://%s.azurewebsites.net/api/report", var.function_app_name)
+    vault_function_app_key_name = var.vault_function_app_key_name
+    key_vault_url               = var.key_vault_url
+  })
 
-  setup_protocol_script = var.setup_protocol ? local.protocol_script : ""
+  smb_protocol_script = var.protocol == "SMB" ? local.setup_smb_protocol_script : ""
+  s3_protocol_script  = var.protocol == "S3" ? local.setup_s3_protocol_script : ""
+  nfs_protocol_script = var.protocol == "NFS" ? local.setup_nfs_protocol_script : ""
+  validation_script   = var.setup_protocol && (var.protocol == "SMB" || var.protocol == "S3") ? local.setup_validation_script : ""
+
+  setup_protocol_script = var.setup_protocol ? compact([local.nfs_protocol_script, local.smb_protocol_script, local.s3_protocol_script]) : []
+
+  custom_data_parts = concat([local.init_script, local.deploy_script, local.validation_script], local.setup_protocol_script)
 
-  custom_data_parts = [
-    local.init_script, local.deploy_script, local.setup_protocol_script
-  ]
   custom_data = join("\n", local.custom_data_parts)
 
   gw_identity_id        = var.vm_identity_name == "" ? azurerm_user_assigned_identity.this[0].id : data.azurerm_user_assigned_identity.this[0].id
@@ -199,7 +211,7 @@ resource "azurerm_linux_virtual_machine" "this" {
   lifecycle {
     ignore_changes = [tags, custom_data]
     precondition {
-      condition     = var.protocol == "NFS" ? var.gateways_number >= 1 : var.gateways_number >= 3 && var.gateways_number <= 8
+      condition     = var.protocol == "NFS" || var.protocol == "S3" ? var.gateways_number >= 1 : var.gateways_number >= 3 && var.gateways_number <= 8
       error_message = "The amount of protocol gateways should be at least 1 for NFS and at least 3 and at most 8 for SMB."
     }
     precondition {

modules/protocol_gateways/setup_s3.sh

@@ -0,0 +1,65 @@
+echo "$(date -u): running s3 script"
+
+wait_for_weka_fs || exit 1
+create_config_fs || exit 1
+
+# wait for weka s3 cluster to be ready in case it was created by another host
+not_ready_hosts=$(weka s3 cluster status | grep 'Not Ready' | wc -l)
+all_hosts=$(weka s3 cluster status | grep 'Host' | wc -l)
+
+function check_cluster_status() {
+  if (( all_hosts > 0 && not_ready_hosts == 0 && all_hosts == cluster_size )); then
+          echo "$(date -u): s3 cluster is already created"
+          weka s3 cluster status
+          exit 0
+  fi
+}
+
+echo "$(date -u): weka S3 cluster does not exist, creating it"
+# get all protocol gateways frontend container ids separated by comma
+all_container_ids_str=$(echo "$all_container_ids" | tr '\n' ',' | sed 's/,$//')
+
+function retry_create_s3_cluster {
+  retry_max=60
+  retry_sleep=30
+  count=$retry_max
+  msg="S3 cluster is created"
+  check_cluster_status
+  while [ $count -gt 0 ]; do
+      weka s3 cluster create $filesystem_name .config_fs --container $all_container_ids_str --port 9000 && break
+      count=$(($count - 1))
+      echo "Retrying create S3 cluster in $retry_sleep seconds..."
+      report "{\"hostname\": \"$HOSTNAME\", \"type\": \"progress\", \"message\": \"$msg\"}"
+      sleep $retry_sleep
+      check_cluster_status && break
+  done
+  [ $count -eq 0 ] && {
+      echo "create S3 cluster command failed after $retry_max attempts"
+      report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"create S3 cluster command failed after $retry_max attempts\"}"
+      echo "$(date -u): create S3 cluster failed"
+      return 1
+  }
+  return 0
+}
+
+if [[ $(weka s3 cluster status) ]]; then
+        check_cluster_status
+        if (( all_hosts > 0 && not_ready_hosts == 0 && all_hosts < cluster_size )); then
+              echo "$(date -u): S3 cluster already exists, adding current container to it"
+              weka s3 cluster containers add $container_id
+              sleep 10s
+              weka s3 cluster status
+              exit 0
+        fi
+else
+  echo "$(date -u): weka S3 cluster does not exist, creating it"
+  retry_create_s3_cluster
+  echo "$(date -u): Successfully create S3 cluster..."
+  weka s3 cluster status
+  weka s3 cluster containers list
+  echo "$(date -u): S3 cluster is created successfully"
+fi
+
+weka s3 cluster status
+
+echo "$(date -u): done running S3 script successfully"

modules/protocol_gateways/setup_smb.sh

@@ -1,136 +1,11 @@
 echo "$(date -u): running smb script"
 weka local ps
 
-# get token for key vault access
-access_token=$(curl 'http://169.254.169.254/metadata/identity/oauth2/token?api-version=2018-02-01&resource=https%3A%2F%2Fvault.azure.net' -H Metadata:true | jq -r '.access_token')
-# get key vault secret
-function_app_key=$(curl "${key_vault_url}secrets/${vault_function_app_key_name}?api-version=2016-10-01" -H "Authorization: Bearer $access_token" | jq -r '.value')
-
-function report {
-  local json_data=$1
-  curl ${report_function_url}?code="$function_app_key" -H 'Content-Type:application/json' -d "$json_data"
-}
-
-function wait_for_weka_fs(){
-  filesystem_name="default"
-  max_retries=30 # 30 * 10 = 5 minutes
-  for (( i=0; i < max_retries; i++ )); do
-    if [ "$(weka fs | grep -c $filesystem_name)" -ge 1 ]; then
-      echo "$(date -u): weka filesystem $filesystem_name is up"
-      break
-    fi
-    echo "$(date -u): waiting for weka filesystem $filesystem_name to be up"
-    sleep 10
-  done
-  if (( i > max_retries )); then
-      err_msg="timeout: weka filesystem $filesystem_name is not up after $max_retries attempts."
-      echo "$(date -u): $err_msg"
-      report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"$err_msg\"}"
-      return 1
-  fi
-}
-
-function create_config_fs(){
-  filesystem_name=".config_fs"
-  size="10GB"
-
-  if [ "$(weka fs | grep -c $filesystem_name)" -ge 1 ]; then
-    echo "$(date -u): weka filesystem $filesystem_name exists"
-    return 0
-  fi
-
-  echo "$(date -u): trying to create filesystem $filesystem_name"
-  output=$(weka fs create $filesystem_name default $size 2>&1)
-  # possiible outputs:
-  # FSId: 1 (means success)
-  # error: The given filesystem ".config_fs" already exists.
-  # error: Not enough available drive capacity for filesystem. requested "10.00 GB", but only "0 B" are free
-  if [ $? -eq 0 ]; then
-    echo "$(date -u): weka filesystem $filesystem_name is created"
-    return 0
-  fi
-
-  if [[ $output == *"already exists"* ]]; then
-    echo "$(date -u): weka filesystem $filesystem_name already exists"
-    break
-  elif [[ $output == *"Not enough available drive capacity for filesystem"* ]]; then
-    err_msg="Not enough available drive capacity for filesystem $filesystem_name for size $size"
-    echo "$(date -u): $err_msg"
-    report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"$err_msg\"}"
-    return 1
-  else
-    echo "$(date -u): output: $output"
-    report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"cannot create weka filesystem $filesystem_name\"}"
-    return 1
-  fi
-}
-
 if [[ ${smbw_enabled} == true ]]; then
   wait_for_weka_fs || exit 1
   create_config_fs || exit 1
 fi
 
-# make sure weka cluster is already up
-max_retries=60
-for (( i=0; i < max_retries; i++ )); do
-  if [ $(weka status | grep 'status: OK' | wc -l) -ge 1 ]; then
-    echo "$(date -u): weka cluster is up"
-    break
-  fi
-  echo "$(date -u): waiting for weka cluster to be up"
-  sleep 30
-done
-if (( i > max_retries )); then
-    err_msg="timeout: weka cluster is not up after $max_retries attempts."
-    echo "$(date -u): $err_msg"
-    report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"$err_msg\"}"
-    exit 1
-fi
-
-cluster_size="${gateways_number}"
-
-current_mngmnt_ip=$(weka local resources | grep 'Management IPs' | awk '{print $NF}')
-# get container id
-for ((i=0; i<20; i++)); do
-  container_id=$(weka cluster container | grep frontend0 | grep ${gateways_name} | grep $current_mngmnt_ip | grep UP | awk '{print $1}')
-  if [ -n "$container_id" ]; then
-      echo "$(date -u): frontend0 container id: $container_id"
-      report "{\"hostname\": \"$HOSTNAME\", \"type\": \"progress\", \"message\": \"frontend0 container $container_id is up\"}"
-      break
-  fi
-  echo "$(date -u): waiting for frontend0 container to be up"
-  sleep 5
-done
-
-if [ -z "$container_id" ]; then
-  err_msg="Failed to get the frontend0 container ID."
-  echo "$(date -u): $err_msg"
-  report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"$err_msg\"}"
-  exit 1
-fi
-
-# wait for all containers to be ready
-max_retries=60
-for (( retry=1; retry<=max_retries; retry++ )); do
-    # get all UP gateway container ids
-    all_container_ids=$(weka cluster container | grep frontend0 | grep ${gateways_name} | grep UP | awk '{print $1}')
-    # if number of all_container_ids < cluster_size, do nothing
-    all_container_ids_number=$(echo "$all_container_ids" | wc -l)
-    if (( all_container_ids_number < cluster_size )); then
-        echo "$(date -u): not all containers are ready - do retry $retry of $max_retries"
-        sleep 20
-    else
-        echo "$(date -u): all containers are ready"
-        break
-    fi
-done
-
-if (( retry > max_retries )); then
-    err_msg="timeout: not all containers are ready after $max_retries attempts."
-    echo "$(date -u): $err_msg"
-    report "{\"hostname\": \"$HOSTNAME\", \"type\": \"error\", \"message\": \"$err_msg\"}"
-    exit 1
-fi
 
 # wait for weka smb cluster to be ready in case it was created by another host
 weka smb cluster wait