ood_core 0.24.2 → 0.26.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '08f64f135dbbf6ea8058870d4339c1669586deed5be368d54dd0ff3ae476d8b0'
-  data.tar.gz: 3f04d4f98590cbaa3ce502fca901e23e67638442b2daa6621bc5fb5c640a91ae
+  metadata.gz: 3ef6cd5a8a206393dcaed9d0121e3646dd6fc3d2b0405992c3fcfa5a745e9489
+  data.tar.gz: 16a5cf6f03ed2be4c563dafb7eaa9e2a04d465a4d8a3cea5d3bb14b634d65868
 SHA512:
-  metadata.gz: 0ca0bdf738b58a61f45c030cc202b3a035d0f33782d7a125d5367fcd63e3fe5f83b70c60a61b772fd3ec5872742632af571c7d9be154845b4836e1fe8659d96c
-  data.tar.gz: 2e7c8f01385fb2553e0ac5774442e3d4b1fca9cee72b1c8fd80b9a4436bcbf297ebce7a0a3e802a7e40fffab1a611c8d02ff74cf19d2c4a627ee5e755ed63581
+  metadata.gz: 563930fc05d105b46fa1f8942294a06b3fd3ffce665ba6aa93fae3a21759613fdf229c9cca4f5ec8bd99c7b57cbb9e7e4348420de47d4d0884a78dfd62c43b5f
+  data.tar.gz: dc88b7bc4a28e96725bcd2caf0e40659bca94a40ab1c50432af0eee2b649f4a48356e858a7f9ceca1a8a18f99c823bf9ffc4c7cecefecb9820d8ce7621c7c51b

data/CHANGELOG.md

@@ -7,10 +7,15 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+# [0.25.0] - 03-27-2024
+
+- [828](https://github.com/OSC/ood_core/pull/828) and [826](https://github.com/OSC/ood_core/pull/826)
+  add configurable heartbeats to noVNC connections.
+
 # [0.24.2] - 01-24-2024
 
 - [823](https://github.com/OSC/ood_core/pull/823) Corrected a mistake in converting duration to seconds.
-* [821](https://github.com/OSC/ood_core/pull/821) add container_start_args to pass options to the start command.
+- [821](https://github.com/OSC/ood_core/pull/821) add container_start_args to pass options to the start command.
 
 ## [0.24.1] - 11-29-2023
 

data/lib/ood_core/batch_connect/templates/vnc.rb

@@ -24,6 +24,9 @@ module OodCore
         # @option context [#to_s] :websockify_cmd
         #   ("${WEBSOCKIFY_CMD:-/opt/websockify/run}") the path to the
         #   websockify script (assumes you don't modify `:after_script`)
+        # @option context [#to_s] :websockify_heartbeat_seconds
+        #   ("${WEBSOCKIFY_HEARTBEAT_SECONDS:-30}") the websockify heartbeat
+        #   duration in seconds. (assumes you don't modify `:after_script`)
         # @option context [#to_s] :vnc_log ("vnc.log") path to vnc server log
         #   file (assumes you don't modify `:before_script` or `:after_script`)
         # @option context [#to_s] :vnc_passwd ("vnc.passwd") path to the file
@@ -130,6 +133,8 @@ module OodCore
           # successful connections so that the password can be reset
           def after_script
             websockify_cmd = context.fetch(:websockify_cmd, "${WEBSOCKIFY_CMD:-/opt/websockify/run}").to_s
+            websockify_hb = context.fetch(:websockify_heartbeat_seconds, "${WEBSOCKIFY_HEARTBEAT_SECONDS:-30}").to_s
+            websockify_timeout_seconds = context.fetch(:websockify_timeout_seconds, '${WEBSOCKIFY_TIMEOUT_SECONDS:-10}').to_s
 
             <<-EOT.gsub(/^ {14}/, "")
               #{super}
@@ -139,9 +144,10 @@ module OodCore
               start_websockify() {
                 local log_file="./websockify.log"
                 # launch websockify in background and redirect all output to a file.
-                #{websockify_cmd} $1 $2 &> $log_file &
+                #{websockify_cmd} $1 --heartbeat=#{websockify_hb} $2 &> $log_file &
                 local ws_pid=$!
                 local counter=0
+                local max_timeout=#{websockify_timeout_seconds}
 
                 # wait till websockify has successfully started
                 echo "[websockify]: pid: $ws_pid (proxying $1 ==> $2)" >&2
@@ -152,9 +158,9 @@ module OodCore
                   if ! ps $ws_pid > /dev/null; then
                     echo "[websockify]: failed to launch!" >&2
                     return 1
-                  elif [ $counter -ge 5 ]; then
-                    # timeout after ~5 seconds
-                    echo "[websockify]: timed-out :(!" >&2
+                  elif [ $counter -ge $max_timeout ]; then
+                    # timeout after max_timeout seconds
+                    echo "[websockify]: timed-out after $max_timeout seconds :(!" >&2
                     return 1
                   else
                     sleep 1

data/lib/ood_core/job/adapter.rb

@@ -213,6 +213,14 @@ module OodCore
       def queues
         []
       end
+
+      # Return the list of nodes for this scheduler.
+      #
+      # Subclasses that do not implement this will return empty arrays.
+      # @return [Array<NodeInfo>]
+      def nodes
+        []
+      end
     end
   end
 end

data/lib/ood_core/job/adapters/slurm.rb

@@ -41,7 +41,7 @@ module OodCore
         # calculated from gres string
         # @return [Integer] the number of gpus in gres
         def self.gpus_from_gres(gres)
-          gres.to_s.scan(/gpu:[^,]*(\d+)/).flatten.map(&:to_i).sum
+          gres.to_s.scan(/gpu[^(,]*[:=](\d+)/).flatten.map(&:to_i).sum
         end
 
         # Object used for simplified communication with a Slurm batch server
@@ -169,6 +169,7 @@ module OodCore
                 # jobs << job
                 #
                 # assuming keys and values are same length! if not we have an error!
+                line = line.encode('UTF-8', invalid: :replace, undef: :replace)
                 values = line.chomp(RECORD_SEPARATOR).strip.split(UNIT_SEPARATOR)
                 jobs << Hash[fields.keys.zip(values)] unless values.empty?
               end
@@ -327,13 +328,37 @@ module OodCore
 
             [].tap do |ret_arr|
               info_raw.each_line do |line|
-                ret_arr << str_to_acct_info(line)
+                ret_arr << str_to_queue_info(line)
               end
             end
           end
 
+          def all_sinfo_node_fields
+            {
+              procs: '%c',
+              name: '%n',
+              features: '%f'
+            }
+          end
+
+          def nodes
+            args = all_sinfo_node_fields.values.join(UNIT_SEPARATOR)
+            output = call('sinfo', '-ho', "#{RECORD_SEPARATOR}#{args}")
+
+            output.each_line(RECORD_SEPARATOR).map do |line|
+              values = line.chomp(RECORD_SEPARATOR).strip.split(UNIT_SEPARATOR)
+
+              next if values.empty?
+
+              data = Hash[all_sinfo_node_fields.keys.zip(values)]
+              data[:name] = data[:name].to_s.split(',').first
+              data[:features] = data[:features].to_s.split(',')
+              NodeInfo.new(**data)
+            end.compact
+          end
+
           private
-            def str_to_acct_info(line)
+            def str_to_queue_info(line)
               hsh = line.split(' ').map do |token|
                 m = token.match(/^(?<key>\w+)=(?<value>.+)$/)
                 [m[:key], m[:value]]
@@ -349,6 +374,7 @@ module OodCore
 
 
               hsh[:deny_accounts] = hsh[:DenyAccounts].nil? ? [] : hsh[:DenyAccounts].to_s.split(',')
+              hsh[:tres] = hsh[:TRES].nil? ? {} : hsh[:TRES].to_s.split(',').map { |str| str.split('=') }.to_h
 
               OodCore::Job::QueueInfo.new(**hsh)
             end
@@ -669,6 +695,10 @@ module OodCore
           @slurm.queues
         end
 
+        def nodes
+          @slurm.nodes
+        end
+
         private
           # Convert duration to seconds
           def duration_in_seconds(time)

data/lib/ood_core/job/adapters/systemd/launcher.rb

@@ -204,7 +204,7 @@ class OodCore::Job::Adapters::LinuxSystemd::Launcher
 
   # List all Systemd sessions on destination_host started by this adapter
   def list_remote_systemd_session(destination_host)
-    cmd = ssh_cmd(destination_host, ['systemctl', '--user', 'show', '-t', 'service', '--state=running', "#{session_name_label}-*"])
+    cmd = ssh_cmd(destination_host, ['systemctl', '--user', 'show', '-t', 'service', '--state=running', "#{session_name_label}-\\*"])
 
     # individual units are separated with an empty line
     call(*cmd).split("\n\n").map do |oneunit|

data/lib/ood_core/job/node_info.rb

@@ -10,17 +10,26 @@ module OodCore
       # @return [Integer, nil] number of procs
       attr_reader :procs
 
+      # The features associated with this node.
+      # @return [Array<String>, []]
+      attr_reader :features
+
       # @param name [#to_s] node name
       # @param procs [#to_i, nil] number of procs
-      def initialize(name:, procs: nil, **_)
+      # @param features [#to_a, []] list of features
+      def initialize(name:, procs: nil, features: [], **_)
         @name  = name.to_s
         @procs = procs && procs.to_i
+        @features = features.to_a
       end
 
       # Convert object to hash
       # @return [Hash] object as hash
       def to_h
-        { name: name, procs: procs }
+        instance_variables.map do |var|
+          name = var.to_s.gsub('@', '').to_sym
+          [name, send(name)]
+        end.to_h
       end
 
       # The comparison operator

data/lib/ood_core/job/queue_info.rb

@@ -20,9 +20,13 @@ class OodCore::Job::QueueInfo
   # The accounts that are not allowed to use this queue.
   attr_reader :deny_accounts
 
+  # An Hash of Trackable Resources and their values.
+  attr_reader :tres
+
   def initialize(**opts)
     @name = opts.fetch(:name, 'unknown')
     @qos = opts.fetch(:qos, [])
+    @tres = opts.fetch(:tres, {})
 
     allow_accounts = opts.fetch(:allow_accounts, nil)
     @allow_accounts = if allow_accounts.nil?
@@ -42,4 +46,8 @@ class OodCore::Job::QueueInfo
       [name, send(name)]
     end.to_h
   end
+
+  def gpu?
+    tres.keys.any? { |name| name.to_s.match?(%r{^gres/gpu($|:)}i) }
+  end
 end

data/lib/ood_core/version.rb

@@ -1,4 +1,4 @@
 module OodCore
   # The current version of {OodCore}
-  VERSION = "0.24.2"
+  VERSION = "0.26.1"
 end

data/ood_core.gemspec

@@ -23,10 +23,10 @@ Gem::Specification.new do |spec|
   spec.required_ruby_version = ">= 2.5.0"
 
   spec.add_runtime_dependency "ood_support", "~> 0.0.2"
-  spec.add_runtime_dependency "ffi", "~> 1.9", ">= 1.9.6"
+  spec.add_runtime_dependency "ffi", "~> 1.16.3"
   spec.add_runtime_dependency "rexml", "~> 3.2"
   spec.add_development_dependency "bundler", "~> 2.1"
-  spec.add_development_dependency "rake", "~> 13.1.0"
+  spec.add_development_dependency "rake", "~> 13.2.0"
   spec.add_development_dependency "rspec", "~> 3.0"
   spec.add_development_dependency "pry", "~> 0.10"
   spec.add_development_dependency "timecop", "~> 0.8"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ood_core
 version: !ruby/object:Gem::Version
-  version: 0.24.2
+  version: 0.26.1
 platform: ruby
 authors:
 - Eric Franz
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2024-01-31 00:00:00.000000000 Z
+date: 2024-07-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ood_support
@@ -32,20 +32,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.9'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.9.6
+        version: 1.16.3
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.9'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.9.6
+        version: 1.16.3
 - !ruby/object:Gem::Dependency
   name: rexml
   requirement: !ruby/object:Gem::Requirement
@@ -80,14 +74,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 13.1.0
+        version: 13.2.0
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 13.1.0
+        version: 13.2.0
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement