anvil/tools/anvil-manage-files

#!/usr/bin/perl
# 
# This handles moving around and managing files on Anvil! nodes, DR hosts and Striker dashboards.
# 
# When this is called (periodically by the daemon of after an upload / ISO generation);
# - 1. Check 'incoming/' for files. For any found, generate an md5sum and see if the file name and sum match 
#      anything in the database from any host;
#      - If so, update/add an entry in 'file_locations'
#      - If not, create a new entry in 'files' and then add the first entry in 'file_locations'
# - 2. Check 'file_locations' for any files on this system, and verify they exist still. 
#      - If not, check the other files for one with a matching md5sum. If found, handle as a rename.
#      - If not found at all, search for the file according to the search rules and copy to here if found.
#      - If not found anywhere, remove it from 'file_locations' and send an alert.
#      - If found, check the size. If it differs, recalculate the md5sum.
# - 3. If called with '--rename --file <filename> --to <newname>', rename the file and update 'files'.
# - 4. If called with '--delete', remove from 'file_locations' and then remove from the local storage. If 
#      also used with '--everywhere', then all copies on all systems we know about will be deleted. This is 
#      done by registering a job against all known hosts. As such, if this is called and the target file 
#      doesn't exist, it just clears the job and then exits.
# - 5. If called with '--is-script=[0|1]', mark as 'script' in the 'files' table and set/remove the executable bit.
#
# Exit codes;
# 0 = Normal exit or md5sum of this program changed and it exited to reload.
# 1 = No databases available.
# 2 = Another process that is still running has picked up this job.
# 3 = No filename specified when needed by another switch.
# 4 = Request to change the file name but the new name wasn't given.
# 5 = The file specified was not found.
# 6 = The file to delete is not under '/mnt/shared/'.
# 
# TODO: 
# - 
# 
# NOTE:
# - 
# 

use strict;
use warnings;
use Anvil::Tools;
use Data::Dumper;


# Disable buffering
$| = 1;

my $THIS_FILE           =  ($0 =~ /^.*\/(.*)$/)[0];
my $running_directory   =  ($0 =~ /^(.*?)\/$THIS_FILE$/)[0];
if (($running_directory =~ /^\./) && ($ENV{PWD}))
{
	$running_directory =~ s/^\./$ENV{PWD}/;
}

my $anvil = Anvil::Tools->new();

$anvil->data->{switches}{'job-uuid'}  = "";
$anvil->data->{switches}{'rename'}    = "";
$anvil->data->{switches}{'is-script'} = "";
$anvil->data->{switches}{file}        = "";
$anvil->data->{switches}{to}          = "";
$anvil->data->{switches}{'delete'}    = "";
$anvil->data->{switches}{everywhere}  = "";
$anvil->Get->switches;
$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { 
	"switches::job-uuid"   => $anvil->data->{switches}{'job-uuid'},
	"switches::rename"     => $anvil->data->{switches}{'rename'},
	"switches::is-script"  => $anvil->data->{switches}{'is-script'},
	"switches::file"       => $anvil->data->{switches}{file},
	"switches::to"         => $anvil->data->{switches}{to},
	"switches::delete"     => $anvil->data->{switches}{'delete'},
	"switches::everywhere" => $anvil->data->{switches}{everywhere},
}});

# Connect or die
$anvil->Database->connect;
$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, level => 2, secure => 0, key => "log_0132"});
if (not $anvil->data->{sys}{database}{connections})
{
	# No databases, exit.
	$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0003"});
	$anvil->nice_exit({exit_code => 1});
}

# If we have a job_uuid, pick it up.
if ($anvil->data->{switches}{'job-uuid'})
{
	my $return = $anvil->Job->get_job_details({check => 1, job_uuid => $anvil->data->{switches}{'job-uuid'}});
	if ($return == 1)
	{
		# It's not a UUID.
		$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 1, list => { 'return' => $return }});
		$anvil->nice_exit({code => 2});
	}
	if ($return == 2)
	{
		# This job is being handled by another process that is still alive.
		$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 1, list => { 'return' => $return }});
		$anvil->nice_exit({code => 3});
	}
	
	# If there's a progress, clear it.
	$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 1, list => { 'job::job_progress' => $anvil->data->{job}{job_progress} }});
	if ($anvil->data->{job}{job_progress})
	{
		$anvil->Job->update_progress({
			progress => 1, 
			message  => "clear", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
	}
}

# What are we doing?
if ($anvil->data->{switches}{'rename'})
{
	handle_rename($anvil);
}
elsif ($anvil->data->{switches}{'delete'}))
{
	handle_delete($anvil);
}
elsif ($anvil->data->{switches}{'is-script'}))
{
	handle_script($anvil);
}


# We're done
$anvil->nice_exit({exit_code => 0});


#############################################################################################################
# Private functions.                                                                                        #
#############################################################################################################

# This gets the file_uuid for a given file name and/or md5sum. If the file isn't found, an empty string is
# returned.
sub get_file_uuid
{
	my ($anvil, $file_md5sum, $file_name) = @_;
	$file_md5sum = "" if not defined $file_md5sum;
	$file_name   = "" if not defined $file_name;
	$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { 
		file_md5sum => $file_md5sum, 
		file_name   => $file_name,
	}});
	
	### TODO: At some point, we'll need to deal with the possibility that the same file name with 
	###       different md5sums might exist in the database.
	# If I have the md5sum, search using that. If I have the filename only, then we'll fall back to that.
	my $query = "
SELECT 
    file_uuid 
FROM 
    files 
WHERE 
";
	if ($file_md5sum)
	{
		$query .= "    file_md5sum = ".$anvil->data->{sys}{database}{use_handle}->quote($file_md5sum)."\n";
	}
	elsif ($file_name)
	{
		$query .= "    file_name = ".$anvil->data->{sys}{database}{use_handle}->quote($file_name)."\n";
	}
	my $results = $anvil->Database->query({query => $query, source => $THIS_FILE, line => __LINE__});
	my $count   = @{$results};
	$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { 
		results => $results, 
		count   => $count, 
	}});
	if (not $count)
	{
		# File wasn't found in the database
		return("");
	}
	my $file_uuid = $results->[0]->[0];
	$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { 
		file_uuid => $file_uuid, 
	}});
	
	return($file_uuid);
}

# This handles toggling a file to marked or unmarked as a script.
sub handle_script
{
	my ($anvil) = @_;
	
	if (not $anvil->data->{switches}{file})
	{
		# Um...
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0055"});
		if ($anvil->data->{jobs}{'job-uuid'})
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0055", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 3});
	}
	
	# Find the file_uuid.
	my ($file_uuid) = get_file_uuid($anvil, "", $anvil->data->{switches}{file});
	$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { file_uuid => $file_uuid }});
	
	# Toggle the executable bits.
	if ($anvil->data->{switches}{'is-script'})
	{
		# Is it already executable?
		if (-x $anvil->data->{switches}{file})
		{
			# Switch it on
			$anvil->Storage->change_mode({target => $anvil->data->{switches}{file}, mode => "a+x"});
		}
		else
		{
			# Already a script.
		}
	}
	else
	{
		# Is it executable?
		if (-x $anvil->data->{switches}{file})
		{
			# Switch it off.
			$anvil->Storage->change_mode({target => $anvil->data->{switches}{file}, mode => "a-x"});
		}
		else
		{
			# Already not a script.
		}
	}
	
	# If we have a file UUID, update the 'file_type
	if ($file_uuid)
	{
		# Load the details.
		my $query = "
SELECT 
    file_name, 
    file_size, 
    file_md5sum, 
    file_type 
FROM 
    files 
WHERE 
    file_uuid = ".$anvil->data->{sys}{database}{use_handle}->quote($file_uuid)." 
;";
		$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { query => $query }});
		
		my $results = $anvil->Database->query({query => $query, source => $file ? $file." -> ".$THIS_FILE : $THIS_FILE, line => $line ? $line." -> ".__LINE__ : __LINE__});
		$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { results => $results }});
		my $file_name   = $results->[0]->[0];
		my $file_size   = $results->[0]->[1];
		my $file_md5sum = $results->[0]->[2]; 
		my $file_type   = $results->[0]->[3]; 
		$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { 
			file_name   => $file_name, 
			file_size   => $file_size, 
			file_md5sum => $file_md5sum, 
			file_type   => $file_type, 
		}});
		if (($file_type eq "script") && (not $anvil->data->{switches}{'is-script'}))
		{
			# Figure out what the file type is and update.
			my $mimetype = mimetype($full_path);
			$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { mimetype => $mimetype }});
			
			my $say_mimetype = "other";
			if ($mimetype =~ /cd-image/)
			{
				$say_mimetype = "iso";
			}
			elsif ($mimetype =~ /rpm$/)
			{
				$say_mimetype = "rpm";
			}
			elsif ($mimetype =~ /disk-image/)
			{
				$say_mimetype = "disk-image";
			}
			$anvil->Log->variables({source => $THIS_FILE, line => __LINE__, level => 2, list => { say_mimetype => $say_mimetype }});
			$anvil->Database->insert_or_update_files({
				file_uuid   => $file_uuid, 
				file_name   => $anvil->data->{switches}{file}, 
				file_size   => $file_size, 
				file_md5sum => $file_md5sum, 
				file_type   => $say_mimetype, 
			});
		}
		elsif ($file_type ne "script") && ($anvil->data->{switches}{'is-script'}))
		{
			# Change the file tpye to "script"
			$anvil->Database->insert_or_update_files({
				file_uuid   => $file_uuid, 
				file_name   => $anvil->data->{switches}{file}, 
				file_size   => $file_size, 
				file_md5sum => $file_md5sum, 
				file_type   => "script", 
			});
		}
	}
	
	return(0);
}

# This handles deleting a file. If the requested deletion target doesn't exist, we'll just clear the 
# database. If that doesn't exist either, we still don't error. This is to handle broad requests to delete a
# file everywhere. If we're asked to delete it everywhere, then we'll register a job against all hosts.
sub handle_delete
{
	my ($anvil) = @_;
	
	if (not $anvil->data->{switches}{file})
	{
		# Um...
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0052"});
		if ($anvil->data->{jobs}{'job-uuid'})
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0052", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 3});
	}
	elsif ($anvil->data->{switches}{file} !~ /^\/mnt\/shared\//)
	{
		# We don't do that here...
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0053", variables => { file => $anvil->data->{switches}{file} }});
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0053,!!file!".$anvil->data->{switches}{file}."!!", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 6});
	}
	
	# Does the file exist?
	if (-e $anvil->data->{switches}{file})
	{
		# Delete it.
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 1, key => "log_0263", variables => { file => $anvil->data->{switches}{file} }});
		
		unlink $anvil->data->{switches}{file};
		
		# Sleep and verify
		sleep 1;
		if (-e $anvil->data->{switches}{file})
		{
			# Failed to delete...
			# TODO:
		}
		else
		{
			# Deleted successfully. Remove from 'file_locations'
			### TODO: Find the file_uuid, then if found, see if we have and entry in file_location and then delete it.
		}
	}
	
	return(0);
}

# This handles renaming files.
sub handle_rename
{
	my ($anvil) = @_;
	
	# Do we have the current file name and the new one?
	if (not $anvil->data->{switches}{file})
	{
		# Um...
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0049"});
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0049", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 3});
	}
	elsif (not $anvil->data->{switches}{to})
	{
		# We need a target
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0050", variables => { file => $anvil->data->{switches}{file} }});
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0050,!!file!".$anvil->data->{switches}{file}."!!", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 4});
	}
	elsif (not -f $anvil->data->{switches}{file})
	{
		# The file to rename doesn't exist.
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0051", variables => { file => $anvil->data->{switches}{file} }});
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0051,!!file!".$anvil->data->{switches}{file}."!!", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 5});
	}
	elsif (-e $anvil->data->{switches}{to})
	{
		# There's already a file (or directory or something) with that name.
		$anvil->Log->entry({source => $THIS_FILE, line => __LINE__, 'print' => 1, level => 0, key => "error_0052", variables => { file => $anvil->data->{switches}{file}, to => $anvil->data->{switches}{to} }});
		$anvil->Job->update_progress({
			progress => 0, 
			message  => "error_0052,!!file!".$anvil->data->{switches}{file}."!!,!!to!".$anvil->data->{switches}{to}."!!", 
			job_uuid => $anvil->data->{jobs}{'job-uuid'},
		});
		$anvil->nice_exit({exit_code => 6});
	}
	
	### TODO: Left off here
	
	return(0);
}