Subversion Repositories havirt

#!/usr/bin/env perl

# Common library for havirt. Basically, just a place to put things which may be used by any

# part of havirt. More for organizations purposes.

# Copyright 2024 Daily Data, Inc.

# 

# Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following 

# conditions are met:

#

#   Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

#   Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer 

#   in the documentation and/or other materials provided with the distribution.

#   Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived

#   from this software without specific prior written permission.

# 

# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT

# NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL

# THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES

# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS

# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE

# OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

# v0.0.1 20240602 RWR

# Initial setup

package havirt;

use warnings;

use strict;  

BEGIN {

   use FindBin;

   use File::Spec;

   # use libraries from the directory this script is in

   use Cwd 'abs_path';

   use File::Basename;

   use lib dirname( abs_path( __FILE__ ) );

}

use Data::Dumper qw(Dumper); # Import the Dumper() subroutine

# define the version number

# see https://metacpan.org/pod/release/JPEACOCK/version-0.97/lib/version.pod

use version;

our $VERSION = version->declare("1.0.0");

use Exporter;

our @ISA = qw( Exporter );

our @EXPORT = qw( 

                  &readDB

                  &writeDB

                  &report

                  &scan

                  &makeCommand

                  &forceScan

                  &executeAndWait

                  &findDomain

                  &diffArray

                  &makeConfig

                  &readConfig

                  &getAvailableResources

                  &resource

                  &validateResources

                  &migrate

                );

# read a DB file (just a YAML)

# if $lock is set, will create a "lock" file so other processes will

# not try to write to it. Using custom code as flock is automagically

# release when the file is read

sub readDB {

   my $lock = shift;

   my $lockFileName = "$main::config->{'status db filename'}.lock";

   my $lockTime = 5; # maximum time to wait for lock to clear

   # wait for lock to clear if it exists, if we are wanting a lock

   # and we have tried it for $locktime iterations

   while ( $lock && -f $lockFileName && $lockTime-- ) {

      sleep 1; # wait one second, then try again

   }

   if ( $lock ) {

      die "Something has $main::config->{'status db filename'} locked, aborting\n" if -f $lockFileName;

      `touch $lockFileName`;

   }

   my $yaml = YAML::Tiny->new( {} );

   if ( -f $main::config->{'status db filename'} ) {

      $yaml = YAML::Tiny->read( $main::config->{'status db filename'} );

   }

   $main::statusDB = $yaml->[0];

}

sub writeDB {

   my $yaml = YAML::Tiny->new( $main::statusDB );

   $yaml->write( $main::config->{'status db filename'} );

   unlink "$main::config->{'status db filename'}.lock" if -f "$main::config->{'status db filename'}.lock"; # release any lock we might have on it

}

sub report {

   if ( $main::config->{'flags'}->{'format'} eq 'tsv' ) {

      return &report_tsv( @_ );

   } else {

      return &report_screen( @_ );

   }

}

sub report_tsv {

   my ( $header, $data ) = @_;

   my @output;

   push @output, join( "\t", @$header );

   for( my $line = 0; $line < @$data; $line++ ) {

      push @output, join( "\t", @{$data->[$line]} );

   } # for

   return join( "\n", @output ) . "\n";

}

sub report_screen {

   my ( $header, $data ) = @_;

   my @output;

   my @widths;

   my $column;

   my $row;

   # First, initialize by using the length of the headers

   for ( $column = 0; $column < @$header; $column++ ) {

      @widths[$column] = length( $header->[$column] );

   }

   # now, go through all data in each row, for each column, and increment the width if it is larger

   for ( $row = 0; $row < @$data; $row++ ) {

      for ( $column = 0; $column < @$header; $column++ ) {

         $widths[$column] = length( $data->[$row][$column] ) 

            if length( $data->[$row][$column] ) > $widths[$column];

      } # for column

   } # for row

   # actually do the print now

   my @format;

   for ( $column = 0; $column < @widths; $column++ ) {

      push ( @format, '%' . $widths[$column] . 's' );

   }

   my $format = join( ' ', @format ) . "\n";

   my $output = sprintf( $format, @$header );

   for ( $row = 0; $row < @$data; $row++ ) {

      $output .= sprintf( $format, @{$data->[$row]} );

   } # for row

   return $output;

}

# scans a node to determine which domains are running on it

sub getDomainsOnNode {

   my $node = shift;

   my $command = &main::makeCommand( $node, 'virsh list' );

   print "havirt.pm:getDomainsOnNode, command is $command\n" if $main::config->{'flags'}->{'debug'} > 2;

   my @nodeList = grep { /^\s*\d/ } `$command`;

   for ( my $i = 0; $i < @nodeList; $i++ ) {

      if ( $nodeList[$i] =~ m/\s*\d+\s*([^ ]+)/ ) {

         $nodeList[$i] = $1;

      }

   }

   my %hash = map{ $_ => time } @nodeList;

   return \%hash;

}

# find node a domain is on

# first parameter is the domain name

# rest of @_ is list of nodes to search

# if no nodes passed in, will search all known nodes

# returns first node found with the domain, or an empty string if not found

# possibly not being used??

sub findDomain {

   my $domainName = shift;

   my @node = @_;

   my $foundNode = '';

   &readDB();

   unless ( @node ) {

      @node = keys %{$main::statusDB->{'node'} };

      print "findDomain, nodes = " . join( "\t", @node ) . "\n" if $main::config->{'flags'}->{'debug'} > 1;

   }

   foreach my $thisNode ( @node ) {

      my $command = &main::makeCommand( $thisNode, 'virsh list' );

      my $output = `$command`;

      print "findDomain, $thisNode list =\n" . $output . "\n" if $main::config->{'flags'}->{'debug'} > 1;;

      return $thisNode if ( $output =~ m/$domainName/ );

   }

   return '';

}

# check one or more nodes and determine which domains are running on them.

# defaults to everything in the node database, but the -t can have it run on only one

# this is the function that should be run every few minutes on one of the servers

sub scan {

   my @targets = @_;

   if ( -f $main::config->{'last scan filename'} && ! $main::config->{'flags'}->{'yes'} ) {

      my $lastScan = time - ( stat( $main::config->{'last scan filename'} ) ) [9];

      return "Scan was run $lastScan seconds ago\n" unless $lastScan > $main::config->{'minum scan time'};

   }

   `touch $main::config->{'last scan filename'}`;

   &main::readDB(1);

   print Dumper( $main::statusDB->{'nodePopulation'} ) if $main::config->{'flags'}->{'debug'} > 2;

   if ( $main::config->{'flags'}->{'target'} ) {

      push @targets, $main::config->{'flags'}->{'target'};

   }

   @targets = keys %{$main::statusDB->{'node'}} unless @targets;

   print "Scanning " . join( "\n", @targets ) . "\n" if $main::config->{'flags'}->{'debug'};

   foreach my $node (@targets) {

      $main::statusDB->{'nodePopulation'}->{$node}->{'running'} = &getDomainsOnNode( $node );

      $main::statusDB->{'nodePopulation'}->{$node}->{'lastchecked'} = time;

      foreach my $domain ( keys %{$main::statusDB->{'nodePopulation'}->{$node}->{'running'}} ) {

         # make sure there is an entry for all of these domains

         $main::statusDB->{'virt'}->{$domain} = {} unless exists( $main::statusDB->{'virt'}->{$domain} );

      }

      print Dumper( $main::statusDB->{'nodePopulation'}->{$node} ) if $main::config->{'flags'}->{'debug'} > 2;

   }

   &main::writeDB();

   return "Node(s) updated\n";

}

# makes the command that will be run on a node

# Created as a sub so we can change format easily

# if node is the node we're on, we don't need to do a remote call

# if node is null, we'll assume we do the command here

# otherwise, we'll do an ssh to the node and run the command there

sub makeCommand {

   my ( $node, $command ) = @_;

   my $me = `hostname`;

   chomp $me;

   if ( ! $node || $node eq $me ) {

      return $command;

   } else {

      return "ssh $node '$command'";

   }

}

# force a node scan, even if time has not expired

sub forceScan {

   my $save = $main::config->{'flags'}->{'yes'};

   $main::config->{'flags'}->{'yes'} = 1;

   &main::scan();

   $main::config->{'flags'}->{'yes'} = $save;

}

# executes command $command, then repeatedly runs virsh list

# on $scanNode, grep'ing for $scanDomain

# $condition is 1 (true) or 0 (false)

sub executeAndWait {

   my ( $command, $scanNode, $scanDomain, $condition ) = @_;

   my $waitSeconds = 5; # number of seconds to wait before checking again

   my $maxIterations = 60 / $waitSeconds; # maximum number of tries

   print "Running [$command], then waiting $waitSeconds to check if complete\n" if $main::config->{'flags'}->{'debug'};

   `$command`;

   my $waitCommand = &makeCommand( $scanNode, "virsh list | grep $scanDomain" );

   my $output = '';

   do {

      return 0 unless ( $maxIterations-- ); # we've waited too long, so probably not working

      print '. ';

      sleep 1;

      $output = `$waitCommand`;

      print "[$waitCommand] returned [$output]\n" if $main::config->{'flags'}->{'debug'} > 1;

   } until ( $condition ? $output : !$output );

   return 1; # made it successful

} 

# find the differences between two arrays (passed by reference)

# first sorts the array, then walks through them one by one

# @$arr1 MUST be larger than @$arr2

sub diffArray {

   my ( $arr1, $arr2 ) = @_;

   my @result;

   @$arr1 = sort @$arr1;

   @$arr2 = sort @$arr2;

   my $i=0;

   my $j=0;

   while ( $i < @$arr1 ) {

      if ( $arr1->[$i] eq $arr2->[$j] ) {

         $i++;

         $j++;

      } elsif ( $arr1->[$i] lt $arr2->[$j] ) {

         push @result, $arr1->[$i];

         $i++;

      } else {

         push @result, $arr2->[$j];

         $j++;

      }

   }

   return \@result;

}

# create a config file if one does not exist

sub makeConfig {

   my ( $config, $filename ) = @_;

   $config->{'script dir'} = $FindBin::RealBin;

   $config->{'script name'} = $FindBin::Script;

   $config->{'db dir'} = $config->{'script dir'} . '/var';

   $config->{'conf dir'} = $config->{'script dir'} . '/conf';

   $config->{'status db filename'} = $config->{'db dir'} . '/status.yaml';

   $config->{'last scan filename'} = $config->{'script dir'} . '/var/lastscan';

   $config->{'minum scan time'} = 5 * 60; # five minutes

   $config->{'node reserved memory'} = 8 * 1024 * 1024; # 8 gigabytes

   $config->{'node reserved vcpu' } = 0; # turn off reserved vcpu

   $config->{'flags'}->{'format'} = 'screen';

   $config->{'flags'}->{'yes'} = 0;

   $config->{'flags'}->{'quiet'} = 0;

   $config->{'flags'}->{'target'} = '';

   $config->{'flags'}->{'dryrun'} = 1;

   $config->{'flags'}->{'debug'} = 0;

   $config->{'flags'}->{'help'} = 0;

   $config->{'flags'}->{'version'} = 0;

   my $yaml = YAML::Tiny->new( $config );

   $yaml->write( $filename );

}

# read the config file and return it

sub readConfig {

   my $filename = shift;

   my $yaml = YAML::Tiny->new( {} );

   if ( -f $filename ) {

      $yaml = YAML::Tiny->read( $filename );

   }

   return $yaml->[0];

}

# find available resource on a node

sub resource {

   my $node = shift;

   die "Can not find node $node in havirt.pm:resource\n"

      unless $main::statusDB->{'node'}->{$node};

   my $return = {

      'memory' => 0,

      'cpu_count' => 0

      };

   foreach my $key ( keys %$return ) {

      $return->{$key} = $main::statusDB->{'node'}->{$node}->{$key}

         if defined $main::statusDB->{'node'}->{$node}->{$key};

   } # foreach

   return $return;

}

sub getAvailableResources {

   my $node = shift;

   &readDB();

   die "Can not find node $node in havirt.pm:resource\n"

      unless $main::statusDB->{'node'}->{$node};

   my $totalResources = &resource( $node );

   print Dumper( $totalResources ) if $main::config->{'flags'}->{'debug'};

   foreach my $domain ( keys %{ $main::statusDB->{'nodePopulation'}->{$node}->{'running'} } ) {

      $totalResources->{'memory'} -= $main::statusDB->{'virt'}->{$domain}->{'memory'};

      $totalResources->{'cpu_count'} -= $main::statusDB->{'virt'}->{$domain}->{'vcpu'};

   }

   return $totalResources;

}

# validate that node has enough resources for the domains which occupy the

# remainder of the stack

# returns 0 on success, or one or more error messages in a string on failure

sub validateResources {

   my $node = shift;

   &readDB();

   my @return;

   my $nodeResources = &getAvailableResources( $node );

   print "In havirt.pm:validateResources, checking if enough room on $node for\n" . join( "\n", @_ ) . "\n"

      if ( $main::config->{'flags'}->{'debug'} );

   # subtract the reserved memory from the node

   $nodeResources->{'memory'} -= $main::config->{'node reserved memory'};

   $nodeResources->{'cpu_count'} -= $main::config->{'node reserved vcpu'} if $main::config->{'node reserved vcpu'};

   while ( my $domain = shift ) {

      $nodeResources->{'memory'} -= $main::statusDB->{'virt'}->{$domain}->{'memory'};

      $nodeResources->{'cpu_count'} -= $main::statusDB->{'virt'}->{$domain}->{'vcpu'};

   }

   print "In havirt.pm:validateResources, $node will have $nodeResources->{memory} memory and $nodeResources->{cpu_count} vcpu's after task\n"

      if ( $main::config->{'flags'}->{'debug'} > 1 );

   push @return, "This action would result in memory of $nodeResources->{memory}" if $nodeResources->{'memory'} <= 0;

   push @return, "This action would result in virtual cpu count of $nodeResources->{cpu_count}" if $nodeResources->{'cpu_count'} <= 0 && $main::config->{'flags'}->{'node reserved vcpu'};

   return @return ? join( "\n", @return ) . "\n" : 0;

}

# migrate domain from current node it is on to $target

sub migrate {

   my ( $virt, $target ) = @_;

   my $return;

   my $node;

   # these are replaced by the safer findDomain

   #&main::forceScan();

   #&main::readDB();

   $node = &main::findDomain( $virt );

   print Dumper( $main::statusDB->{'nodePopulation'} ) if $main::config->{'flags'}->{'debug'} > 2;

   die "I can not find $virt on any node\n" unless $node;

   die "Domain $virt in maintenance mode, can not migrate it\n" if $main::statusDB->{'virt'}->{$virt}->{'maintenance'};

   die "Node $target in maintenance mode, can not migrate anything to it\n" if $main::statusDB->{'node'}->{$target}->{'maintenance'};

   die "$virt already on $target\n" if $target eq $node;

   my $command = &main::makeCommand( $node, "virsh migrate --live --persistent --verbose  $virt qemu+ssh://$target/system" );

   if ( $main::config->{'flags'}->{'yes'} ) { # they want us to actually do it

      $return = ( &main::executeAndWait( $command, $node, $virt, 0 ) ? 'Success' : 'Time Out waiting for shutdown');

      &main::forceScan();

   } else {

      $return = $command;

   }

   return "$return\n";

}