update EC2 image for Julia 1.11.2 and minor changes

Author: decarvalhojunior-fh

CCconfig.toml

@@ -18,7 +18,7 @@ mpiflags = ""
 
 [ec2]
 
-imageid = "ami-0ce455c31c5fec661"  # found at us-east-1 (North Virginia). To use in other regions, copy it.
+imageid = "ami-09121cfdb459a0804"  # found at us-east-1 (North Virginia). To use in other regions, copy it.
 
 # placement_group = "pg-XXXXXXXXXXXX" or "automatic"
 # security_group_id = "sg-XXXXXXXXXXXX" or "automatic"

src/cluster_providers/ec2/ec2_backend.jl

@@ -173,12 +173,12 @@ function ec2_set_up_ssh_connection(cluster_name)
 
     internal_key_name = cluster_name
 
-    if !isdir(joinpath(homedir(),".ssh"))
-        mkdir(joinpath(homedir(),".ssh"))
-    end
+    ssh_path = joinpath(homedir(), ".ssh")
+
+    !isdir(ssh_path) && mkdir(ssh_path)
 
-    keypath = joinpath(homedir(), ".ssh", "$internal_key_name.key")
-    pubpath = joinpath(homedir(), ".ssh", "$internal_key_name.key.pub")
+    keypath = joinpath(ssh_path, "$internal_key_name.key")
+    pubpath = joinpath(ssh_path, "$internal_key_name.key.pub")
 
    # Criar chave interna pública e privada do SSH.
    # chars = ['a':'z'; 'A':'Z'; '0':'9']
@@ -276,10 +276,11 @@ function ec2_create_params(cluster::PeerWorkers, user_data_base64)
 end
 
 function ec2_remove_temp_files(internal_key_name)
-    keypath = joinpath(homedir(), ".ssh", "$internal_key_name.key")
-    pubpath = joinpath(homedir(), ".ssh", "$internal_key_name.key.pub")
-    run(`rm $keypath`)
-    run(`rm $pubpath`)
+    ssh_path = joinpath(homedir(), ".ssh")
+    keypath = joinpath(ssh_path, "$internal_key_name.key")
+    pubpath = joinpath(ssh_path, "$internal_key_name.key.pub")
+    rm(keypath)
+    rm(pubpath)
 end
 
 
@@ -310,7 +311,8 @@ function ec2_set_hostfile(cluster_nodes, internal_key_name)
         end
     end
 
-    keypath = joinpath(homedir(), ".ssh", "$internal_key_name.key")
+    ssh_path = joinpath(homedir(), ".ssh")
+    keypath = joinpath(ssh_path, "$internal_key_name.key")
 
     # Atualiza o hostname e o hostfile.
     for instance in keys(cluster_nodes)
@@ -437,7 +439,7 @@ function ec2_await_status(cluster_nodes, status)
         print("Waiting for $nodeid to be $status ...")
         while ec2_get_instance_status(cluster_nodes[nodeid]) != status
             print(".")
-            sleep(5)
+            sleep(2)
         end
         println("successfull")
     end
@@ -448,7 +450,7 @@ function ec2_await_check(cluster_nodes, status)
         print("Waiting for $nodeid to be $status ...")
         while ec2_get_instance_check(cluster_nodes[nodeid]) != status
             print(".")
-            sleep(5)
+            sleep(2)
         end
         println("successfull")
     end
@@ -528,7 +530,7 @@ function ec2_create_mount_point(file_system_id, subnet_id, security_group_id)
     status = Efs.describe_file_systems(Dict("FileSystemId" => file_system_id))["FileSystems"][1]["LifeCycleState"]
     while status != "available"
         println("Waiting for File System to be available...")
-        sleep(5)
+        sleep(2)
         status = Efs.describe_file_systems(Dict("FileSystemId" => file_system_id))["FileSystems"][1]["LifeCycleState"]
     end
     println("Creating Mount Target...")
@@ -537,7 +539,7 @@ function ec2_create_mount_point(file_system_id, subnet_id, security_group_id)
     status = Efs.describe_mount_targets(Dict("MountTargetId" => mount_target_id))["MountTargets"][1]["LifeCycleState"]
     while status != "available"
         println("Waiting for mount target to be available...")
-        sleep(5)
+        sleep(2)
         status = Efs.describe_mount_targets(Dict("MountTargetId" => mount_target_id))["MountTargets"][1]["LifeCycleState"]
     end
     mount_target_id
@@ -555,7 +557,7 @@ function ec2_delete_efs(file_system_id)
     end
     while length(Efs.describe_mount_targets(Dict("FileSystemId" => file_system_id))["MountTargets"]) != 0
         println("Waiting for mount targets to be deleted...")
-        sleep(5)
+        sleep(2)
     end
     Efs.delete_file_system(file_system_id)
 end
@@ -580,7 +582,8 @@ ec2_can_resume(cluster::Cluster) = ec2_cluster_status(cluster, ["stopped"])
 # If some instance is not in "interrupted" or "running" state, raise an exception.
 # PUBLIC
 function ec2_resume_cluster(cluster::Cluster)
-    keypath = joinpath(homedir(), ".ssh", "$(cluster.name).key")
+    ssh_path = joinpath(homedir(), ".ssh")
+    keypath = joinpath(ssh_path, "$(cluster.name).key")
 
     ec2_start_instances(cluster)
     ec2_await_status(cluster.cluster_nodes, "running")

src/cluster_providers/gcp/gcp_backend.jl

@@ -81,7 +81,7 @@ function gcp_terminate_cluster(cluster::Cluster)
         status = gcp_get_instance_status(instance[2])
         while status != "terminated"
             println("Waiting for instances to terminate...")
-            sleep(5)
+            sleep(2)
             status = gcp_get_instance_status(instance[2])
         end
     end
@@ -405,7 +405,7 @@ function gcp_await_status(cluster_nodes, status)
         print("Waiting for $nodeid to be $status ...")
         while gcp_get_instance_status(cluster_nodes[nodeid]) != status
             print(".")
-            sleep(5)
+            sleep(2)
         end
         println("successfull")
     end
@@ -416,7 +416,7 @@ function gcp_await_check(cluster_nodes, status)
         print("Waiting for $nodeid to be $status ...")
         while gcp_get_instance_check(cluster_nodes[nodeid]) != status
             print(".")
-            sleep(5)
+            sleep(2)
         end
         println("successfull")
     end

src/deploy.jl

@@ -383,6 +383,7 @@ function cluster_interrupt(cluster_handle)
         cluster_type = cluster_features[:cluster_type]
         try
             kill_processes(cluster_handle, cluster_type, cluster_features)
+            sleep(1)
         finally
             interrupt_cluster(node_provider, cluster_handle)
         end