]> git.openstreetmap.org Git - chef.git/blobdiff - cookbooks/hardware/recipes/default.rb
Merge remote-tracking branch 'github/pull/713'
[chef.git] / cookbooks / hardware / recipes / default.rb
index d7c508d9aadabf7dd81929ec698643f1088cd6e8..69afc73ffbee9798316be45bbd0c643381f87e5f 100644 (file)
@@ -19,7 +19,6 @@
 
 include_recipe "apt"
 include_recipe "git"
-include_recipe "munin"
 include_recipe "prometheus"
 include_recipe "sysfs"
 include_recipe "tools"
@@ -226,6 +225,10 @@ else
   end
 end
 
+watchdog_module = %w[hpwdt sp5100_tco].find do |module_name|
+  node[:hardware][:pci]&.any? { |_, pci| pci[:modules]&.any?(module_name) }
+end
+
 if node[:kernel][:modules].include?("ipmi_si")
   package "ipmitool"
   package "freeipmi-tools"
@@ -246,6 +249,8 @@ if node[:kernel][:modules].include?("ipmi_si")
     options "--config.file=/etc/prometheus/ipmi_local.yml"
     subscribes :restart, "template[/etc/prometheus/ipmi_local.yml]"
   end
+
+  watchdog_module ||= "ipmi_watchdog"
 end
 
 package "irqbalance"
@@ -326,6 +331,8 @@ if node[:virtualization][:role] != "guest" ||
   end
 end
 
+include_recipe "apt::hwraid" unless status_packages.empty?
+
 %w[ssacli lsiutil sas2ircu megactl megacli arcconf].each do |tools_package|
   if tools_packages.include?(tools_package)
     package tools_package
@@ -354,8 +361,6 @@ else
   end
 end
 
-include_recipe "apt::hwraid" unless status_packages.empty?
-
 %w[cciss-vol-status mpt-status sas2ircu-status megaclisas-status aacraid-status].each do |status_package|
   if status_packages.include?(status_package)
     package status_package
@@ -405,15 +410,15 @@ intel_nvmes = nvmes.select { |pci| pci[:vendor_name] == "Intel Corporation" }
 if !intel_ssds.empty? || !intel_nvmes.empty?
   package "unzip"
 
-  sst_tool_version = "1.3"
-  sst_package_version = "#{sst_tool_version}.208-0"
+  sst_tool_version = "2-0"
+  sst_package_version = "2.0.300-0"
 
-  # remote_file "#{Chef::Config[:file_cache_path]}/SST_CLI_Linux_#{sst_tool_version}.zip" do
-  #   source "https://downloadmirror.intel.com/743764/SST_CLI_Linux_#{sst_tool_version}.zip"
-  end
+  remote_file "#{Chef::Config[:file_cache_path]}/sst-cli-linux-deb--#{sst_tool_version}.zip" do
+    source "https://sdmsdfwdriver.blob.core.windows.net/files/kba-gcc/drivers-downloads/ka-00085/sst--#{sst_tool_version}/sst-cli-linux-deb--#{sst_tool_version}.zip"
+  end
 
-  execute "#{Chef::Config[:file_cache_path]}/SST_CLI_Linux_#{sst_tool_version}.zip" do
-    command "unzip SST_CLI_Linux_#{sst_tool_version}.zip sst_#{sst_package_version}_amd64.deb"
+  execute "#{Chef::Config[:file_cache_path]}/sst-cli-linux-deb--#{sst_tool_version}.zip" do
+    command "unzip sst-cli-linux-deb--#{sst_tool_version}.zip sst_#{sst_package_version}_amd64.deb"
     cwd Chef::Config[:file_cache_path]
     user "root"
     group "root"
@@ -439,38 +444,21 @@ disks = disks.map do |disk|
     if controller && controller[:device]
       device = controller[:device].sub("/dev/", "")
       smart = disk[:smart_device]
-
-      if device.start_with?("cciss/") && smart =~ /^cciss,(\d+)$/
-        array = node[:hardware][:disk][:arrays][disk[:arrays].first]
-        munin = "cciss-3#{array[:wwn]}-#{Regexp.last_match(1)}"
-      elsif smart =~ /^.*,(\d+)$/
-        munin = "#{device}-#{Regexp.last_match(1)}"
-      elsif smart =~ %r{^.*,(\d+)/(\d+)$}
-        munin = "#{device}-#{Regexp.last_match(1)}:#{Regexp.last_match(2)}"
-      end
     elsif disk[:device]
       device = disk[:device].sub("/dev/", "")
       smart = disk[:smart_device]
-
-      if smart =~ /^.*,(\d+),(\d+),(\d+)$/
-        munin = "#{device}-#{Regexp.last_match(1)}:#{Regexp.last_match(2)}:#{Regexp.last_match(3)}"
-      end
     end
   elsif disk[:device] =~ %r{^/dev/(nvme\d+)n\d+$}
     device = Regexp.last_match(1)
-    munin = device
   elsif disk[:device]
     device = disk[:device].sub("/dev/", "")
-    munin = device
   end
 
-  next if device.nil? || munin.nil?
+  next if device.nil?
 
   Hash[
     :device => device,
-    :smart => smart,
-    :munin => munin,
-    :hddtemp => munin.tr("-:", "_")
+    :smart => smart
   ]
 end
 
@@ -530,48 +518,12 @@ if disks.count.positive?
     private_users false
     protect_clock false
   end
-
-  # Don't try and do munin monitoring of disks behind
-  # an Areca controller as they only allow one thing to
-  # talk to the controller at a time and smartd will
-  # throw errors if it clashes with munin
-  disks = disks.reject { |disk| disk[:smart]&.start_with?("areca,") }
-
-  disks.each do |disk|
-    munin_plugin "smart_#{disk[:munin]}" do
-      target "smart_"
-      conf "munin.smart.erb"
-      conf_variables :disk => disk
-    end
-  end
 else
   service "smartd" do
     action [:stop, :disable]
   end
 end
 
-if disks.count.positive?
-  munin_plugin "hddtemp_smartctl" do
-    conf "munin.hddtemp.erb"
-    conf_variables :disks => disks
-  end
-else
-  munin_plugin "hddtemp_smartctl" do
-    action :delete
-    conf "munin.hddtemp.erb"
-  end
-end
-
-plugins = Dir.glob("/etc/munin/plugins/smart_*").map { |p| File.basename(p) } -
-          disks.map { |d| "smart_#{d[:munin]}" }
-
-plugins.each do |plugin|
-  munin_plugin plugin do
-    action :delete
-    conf "munin.smart.erb"
-  end
-end
-
 if File.exist?("/etc/mdadm/mdadm.conf")
   mdadm_conf = edit_file "/etc/mdadm/mdadm.conf" do |line|
     line.gsub!(/^MAILADDR .*$/, "MAILADDR admins@openstreetmap.org")
@@ -609,19 +561,30 @@ node[:hardware][:blacklisted_modules].each do |module_name|
   end
 end
 
-if node[:hardware][:watchdog]
-  package "watchdog"
+if watchdog_module
+  kernel_module watchdog_module do
+    action :install
+  end
+
+  execute "systemctl-reload" do
+    action :nothing
+    command "systemctl daemon-reload"
+    user "root"
+    group "root"
+  end
 
-  template "/etc/default/watchdog" do
-    source "watchdog.erb"
+  directory "/etc/systemd/system.conf.d" do
     owner "root"
     group "root"
-    mode "644"
-    variables :module => node[:hardware][:watchdog]
+    mode "755"
   end
 
-  service "watchdog" do
-    action [:enable, :start]
+  template "/etc/systemd/system.conf.d/watchdog.conf" do
+    source "watchdog.conf.erb"
+    owner "root"
+    group "root"
+    mode "644"
+    notifies :run, "execute[systemctl-reload]"
   end
 end