Processing/bin/downloadSample.sh

#!/bin/bash
#---------------------------------------------------------------------------------------------------
# Download a list of files
#---------------------------------------------------------------------------------------------------

# Read the arguments
echo ""
echo " downloadSample.sh  $*"
echo ""
dataDir=$1;      shift
book=$1;         shift
dataset=$1;      shift
target=$1;       shift
nCopyProcs=$1;   shift
condorOutput=$1; shift
onlyMissing=$1;  shift

DN=`grid-proxy-info -subject`

# Prepare environment
echo " "
echo " Process dataset: $dataset  of book: $book"
echo "   in directory : $dataDir"
echo "   to target    : $target"
echo "   n copy procs : $nCopyProcs"
echo "   condor output: $condorOutput"
echo "   only missing : $onlyMissing"

mkdir -p $condorOutput/$book/$dataset
makedir --exe  $target
makedir --exe  $target/$book
makedir --exe  $target/$book/$dataset
script=`which downloadFiles.sh`

# cleanup our lists and remake a clean one
#echo "rm -f $condorOutput/$book/$dataset/fileList*.$$.txt*"
rm -f $condorOutput/$book/$dataset/fileList*.$$.txt*

# make list of all local files
if [ "`echo $HOSTNAME | grep mit.edu`" != "" ] && \
 ( [ "`echo $dataDir  | grep /castor/cern.ch`" != "" ] || \
   [ "`echo $target   | grep /castor/cern.ch`" != "" ] )
then
  opt="--simple"
else
  opt=""
fi

list $opt $dataDir/$book/$dataset | sort > $condorOutput/$book/$dataset/fileList-all.$$.txt-bak

# Make sure there are kerberos and globus tickets available
id=`id -u`
mkdir             -p  ~/.krb5/
cp /tmp/x509up_u${id} ~/.krb5/
KRB5CCNAME=`klist -5 | grep 'Ticket cache:' | cut -d' ' -f 3`
if ! [ -z $KRB5CCNAME ]
then
  mkdir    -p  ~/.krb5/
  chmod 0      ~/.krb5
  chmod u=rwx  ~/.krb5
  file=`echo $KRB5CCNAME | cut -d: -f2`
  if [ -f "$file" ]
  then
    cp $file ~/.krb5/ticket
  else
    echo " ERROR -- missing kerberos ticket ($KRB5CCNAME)."
    exit 1
  fi
else
  echo " ERROR -- missing kerberos ticket ($KRB5CCNAME)."
fi

# make list of all remote files
rm -f $condorOutput/$book/$dataset/fileList-all.$$.txt
touch $condorOutput/$book/$dataset/fileList-all.$$.txt

cat   $condorOutput/$book/$dataset/fileList-all.$$.txt-bak | grep root | sort | \
while read line
do
  size=`echo $line | tr -s ' ' | cut -d ' ' -f 1`
  file=`echo $line | tr -s ' ' | cut -d ' ' -f 2`
  file=`basename $file`
  echo "$size $file" >> $condorOutput/$book/$dataset/fileList-all.$$.txt
done

# make list of all local files
if [ "`echo $HOSTNAME | grep mit.edu`" != "" ] && \
    ( [ "`echo $dataDir | grep /castor/cern.ch`" != "" ] || \
      [ "`echo $target | grep /castor/cern.ch`" != "" ] )
then
  opt="--simple"
else
  opt=""
fi

list $opt $target/$book/$dataset | grep root | sort \
     > $condorOutput/$book/$dataset/fileList-done.$$.txt

diff -y $condorOutput/$book/$dataset/fileList-all.$$.txt  \
        $condorOutput/$book/$dataset/fileList-done.$$.txt > diff.$$
echo ""
echo " Files different in size: "
grep \| diff.$$
echo ""
echo " Files available in all and not done: "
grep \< diff.$$
echo ""
echo " Files done but not listed in all available: "
grep \> diff.$$
echo ""
rm      diff.$$

# make list of missing files
rm -f $condorOutput/$book/$dataset/fileList.$$.txt
touch $condorOutput/$book/$dataset/fileList.$$.txt

cat   $condorOutput/$book/$dataset/fileList-all.$$.txt | grep root | \
while read line
do
  size=`echo $line | tr -s ' ' | cut -d ' ' -f 1`
  file=`echo $line | tr -s ' ' | cut -d ' ' -f 2`
  exists=`grep "$file" $condorOutput/$book/$dataset/fileList-done.$$.txt`
  if [ "$exists" == "" ]
  then
    echo "   -missing-- $file with $size bytes"
    echo "$size $file" >> $condorOutput/$book/$dataset/fileList.$$.txt
  # else
  #   echo "   -exists--- $file with $size bytes - exists"
  else
    # now check that size matches
    test=`grep "$size $file" $condorOutput/$book/$dataset/fileList-done.$$.txt`
    if [ "$test" == "" ]
    then
      if [ "$onlyMissing" == "" ]
      then
        echo "   -fileSize- $exists (remote: $size)"
        echo "$size $file" >> $condorOutput/$book/$dataset/fileList.$$.txt
      fi
    fi
  fi
done

nAll=`wc -l $condorOutput/$book/$dataset/fileList-all.$$.txt | cut -d ' ' -f1`
nMissing=`wc -l $condorOutput/$book/$dataset/fileList.$$.txt | cut -d ' ' -f1`
nDone=`wc -l $condorOutput/$book/$dataset/fileList-done.$$.txt | cut -d ' ' -f1`
echo ""
echo " Download Summary "
echo "   All       $nAll"
echo "   Done      $nDone"
echo "   Missing   $nMissing"
echo ""

# construct our job
nFiles=`wc -l $condorOutput/$book/$dataset/fileList.$$.txt | cut -d ' ' -f1`
if   [ "$nFiles" == "" ] || [ "$nFiles" == "0" ]
then
  echo " "
  echo " No more files to download. EXIT."
  exit 0
elif [ $nFiles -lt $nCopyProcs ]
then
  nCopyProcs=$nFiles
fi
# how many files per job?
nFilesPerJob=$(( $nFiles/$nCopyProcs ))
nFilesTmp=$(( $nFilesPerJob*$nCopyProcs ))
if [ $nFilesPerJob == 1 ] && [ $nFiles -gt $nCopyProcs ]
then
  nFilesPerJob=2
elif [ $nFilesTmp -lt $nFiles ]
then
  nFilesPerJob=$(( $nFilesPerJob+1 ))
fi

echo "   n files to copy: $nFiles"
echo "   n files/proc   : $nFilesPerJob"

i=1
next=1
last=$nFilesPerJob

# make sure condor is properly setup for us
if ! [ -z $CONDOR_LOCATION ]
then
  unset  CONDOR_LOCATION
  export CONDOR_CONFIG=/usr/local/condor/etc/condor_config
fi

# stage in the missing files if it is at CERN
if [ "`echo $dataDir | grep /castor/cern.ch`" != "" ]
then
  echo "  scp $condorOutput/$book/$dataset/fileList.$$.txt  $TICKET_HOLDER@lxplus.cern.ch:"
  scp $condorOutput/$book/$dataset/fileList.$$.txt  $TICKET_HOLDER@lxplus.cern.ch:
  echo "  ssh $TICKET_HOLDER@lxplus.cern.ch ./stageSample.py --dataDir=$dataDir/$book/$dataset --fileList=fileList.$$.txt"
  ssh $TICKET_HOLDER@lxplus.cern.ch ./stageSample.py --dataDir=$dataDir/$book/$dataset --fileList=fileList.$$.txt
  echo "  ssh $TICKET_HOLDER@lxplus.cern.ch rm fileList.$$.txt"
  ssh $TICKET_HOLDER@lxplus.cern.ch rm fileList.$$.txt
fi

# loop over the condor jobs and submit them
while [ $i -le $nCopyProcs ] && [ $last -le $nFiles ]
do
  if [ $i == $nCopyProcs ]
  then
    last=$nFiles
  fi

  # say what we are going to submit
  echo "   downloadFiles.sh $dataDir $book $dataset $target $condorOutput $$ $next $last"

  logFile=`echo download:$book/$dataset/${next}-${last}.txt | tr '/' '+'`
  logFile=/tmp/$logFile
  rm -f $logFile

  # prepare the condor_submit files
  cat > submit_$$.cmd <<EOF
Universe                = vanilla
Requirements            = ( (Arch == "INTEL") && (Disk >= DiskUsage) && ((Memory * 1024) >= ImageSize) && (HasFileTransfer) )
Notify_user             = $TICKET_HOLDER@mit.edu
Notification            = Error
Executable              = $script
Arguments               = $dataDir $book $dataset $target $condorOutput $$ $next $last
Rank                    = Mips
GetEnv                  = True
Input                   = /dev/null
Output                  = $condorOutput/$book/$dataset/${next}-${last}.out
Error                   = $condorOutput/$book/$dataset/${next}-${last}.err
Log                     = $logFile
should_transfer_files   = YES
when_to_transfer_output = ON_EXIT

+AccountingGroup        = "group_cmsuser.cmsu0284"

Queue
EOF

#+x509userproxysubject   = $DN

  # submit the jobs
  condor_submit submit_$$.cmd >& /dev/null #>& lastSub
  #cat submit_$$.cmd
  rm  submit_$$.cmd

  # update counters
  next=$(( $next + $nFilesPerJob ))
  last=$(( $last + $nFilesPerJob ))
  i=$(( $i + 1 ))
done

exit 0
Revision:	1.13
Committed:	Thu Aug 9 21:16:03 2012 UTC (12 years, 8 months ago) by paus
Content type:	application/x-sh
Branch:	MAIN
CVS Tags:	Mit_032, Mit_031, Mit_030, Mit_029c, Mit_029b, Mit_030_pre1, Mit_029a, Mit_029, HEAD
Changes since 1.12:	+0 -2 lines
Log Message:	For version 029.
#	User	Rev	Content
1	paus	1.1	#!/bin/bash
2			#---------------------------------------------------------------------------------------------------
3			# Download a list of files
4			#---------------------------------------------------------------------------------------------------
5
6			# Read the arguments
7			echo ""
8	paus	1.11	echo " downloadSample.sh $*"
9	paus	1.1	echo ""
10			dataDir=$1; shift
11			book=$1; shift
12			dataset=$1; shift
13			target=$1; shift
14			nCopyProcs=$1; shift
15			condorOutput=$1; shift
16	paus	1.3	onlyMissing=$1; shift
17
18			DN=`grid-proxy-info -subject`
19	paus	1.1
20			# Prepare environment
21			echo " "
22			echo " Process dataset: $dataset of book: $book"
23			echo " in directory : $dataDir"
24			echo " to target : $target"
25			echo " n copy procs : $nCopyProcs"
26			echo " condor output: $condorOutput"
27	paus	1.3	echo " only missing : $onlyMissing"
28	paus	1.1
29	paus	1.8	mkdir -p $condorOutput/$book/$dataset
30	paus	1.9	makedir --exe $target
31			makedir --exe $target/$book
32	paus	1.8	makedir --exe $target/$book/$dataset
33	paus	1.1	script=`which downloadFiles.sh`
34
35	paus	1.3	# cleanup our lists and remake a clean one
36	paus	1.11	#echo "rm -f $condorOutput/$book/$dataset/fileList.$$.txt"
37			rm -f $condorOutput/$book/$dataset/fileList.$$.txt
38	paus	1.1
39	paus	1.3	# make list of all local files
40	paus	1.4	if [ "`echo $HOSTNAME \| grep mit.edu`" != "" ] && \
41	paus	1.12	( [ "`echo $dataDir \| grep /castor/cern.ch`" != "" ] \|\| \
42			[ "`echo $target \| grep /castor/cern.ch`" != "" ] )
43	paus	1.1	then
44	paus	1.3	opt="--simple"
45	paus	1.1	else
46	paus	1.3	opt=""
47	paus	1.1	fi
48
49	paus	1.11	list $opt $dataDir/$book/$dataset \| sort > $condorOutput/$book/$dataset/fileList-all.$$.txt-bak
50	paus	1.3
51	paus	1.11	# Make sure there are kerberos and globus tickets available
52	paus	1.3	id=`id -u`
53	paus	1.11	mkdir -p ~/.krb5/
54	paus	1.3	cp /tmp/x509up_u${id} ~/.krb5/
55	paus	1.1	KRB5CCNAME=`klist -5 \| grep 'Ticket cache:' \| cut -d' ' -f 3`
56			if ! [ -z $KRB5CCNAME ]
57			then
58	paus	1.12	mkdir -p ~/.krb5/
59	paus	1.1	chmod 0 ~/.krb5
60			chmod u=rwx ~/.krb5
61			file=`echo $KRB5CCNAME \| cut -d: -f2`
62			if [ -f "$file" ]
63			then
64			cp $file ~/.krb5/ticket
65			else
66			echo " ERROR -- missing kerberos ticket ($KRB5CCNAME)."
67			exit 1
68			fi
69			else
70			echo " ERROR -- missing kerberos ticket ($KRB5CCNAME)."
71			fi
72
73			# make list of all remote files
74	paus	1.11	rm -f $condorOutput/$book/$dataset/fileList-all.$$.txt
75			touch $condorOutput/$book/$dataset/fileList-all.$$.txt
76
77			cat $condorOutput/$book/$dataset/fileList-all.$$.txt-bak \| grep root \| sort \| \
78	paus	1.1	while read line
79			do
80			size=`echo $line \| tr -s ' ' \| cut -d ' ' -f 1`
81			file=`echo $line \| tr -s ' ' \| cut -d ' ' -f 2`
82			file=`basename $file`
83	paus	1.11	echo "$size $file" >> $condorOutput/$book/$dataset/fileList-all.$$.txt
84	paus	1.1	done
85
86			# make list of all local files
87	paus	1.9	if [ "`echo $HOSTNAME \| grep mit.edu`" != "" ] && \
88			( [ "`echo $dataDir \| grep /castor/cern.ch`" != "" ] \|\| \
89			[ "`echo $target \| grep /castor/cern.ch`" != "" ] )
90	paus	1.1	then
91	paus	1.3	opt="--simple"
92			else
93			opt=""
94	paus	1.1	fi
95	paus	1.3
96	paus	1.12	list $opt $target/$book/$dataset \| grep root \| sort \
97	paus	1.11	> $condorOutput/$book/$dataset/fileList-done.$$.txt
98
99			diff -y $condorOutput/$book/$dataset/fileList-all.$$.txt \
100			$condorOutput/$book/$dataset/fileList-done.$$.txt > diff.$$
101			echo ""
102			echo " Files different in size: "
103			grep \\| diff.$$
104			echo ""
105			echo " Files available in all and not done: "
106			grep \< diff.$$
107			echo ""
108			echo " Files done but not listed in all available: "
109			grep \> diff.$$
110			echo ""
111			rm diff.$$
112	paus	1.1
113			# make list of missing files
114	paus	1.11	rm -f $condorOutput/$book/$dataset/fileList.$$.txt
115			touch $condorOutput/$book/$dataset/fileList.$$.txt
116
117			cat $condorOutput/$book/$dataset/fileList-all.$$.txt \| grep root \| \
118	paus	1.1	while read line
119			do
120			size=`echo $line \| tr -s ' ' \| cut -d ' ' -f 1`
121			file=`echo $line \| tr -s ' ' \| cut -d ' ' -f 2`
122	paus	1.11	exists=`grep "$file" $condorOutput/$book/$dataset/fileList-done.$$.txt`
123	paus	1.1	if [ "$exists" == "" ]
124			then
125	paus	1.3	echo " -missing-- $file with $size bytes"
126	paus	1.11	echo "$size $file" >> $condorOutput/$book/$dataset/fileList.$$.txt
127	paus	1.1	# else
128	paus	1.3	# echo " -exists--- $file with $size bytes - exists"
129			else
130			# now check that size matches
131	paus	1.11	test=`grep "$size $file" $condorOutput/$book/$dataset/fileList-done.$$.txt`
132	paus	1.3	if [ "$test" == "" ]
133			then
134			if [ "$onlyMissing" == "" ]
135			then
136			echo " -fileSize- $exists (remote: $size)"
137	paus	1.11	echo "$size $file" >> $condorOutput/$book/$dataset/fileList.$$.txt
138	paus	1.3	fi
139			fi
140	paus	1.1	fi
141			done
142	paus	1.11
143			nAll=`wc -l $condorOutput/$book/$dataset/fileList-all.$$.txt \| cut -d ' ' -f1`
144			nMissing=`wc -l $condorOutput/$book/$dataset/fileList.$$.txt \| cut -d ' ' -f1`
145			nDone=`wc -l $condorOutput/$book/$dataset/fileList-done.$$.txt \| cut -d ' ' -f1`
146	paus	1.10	echo ""
147			echo " Download Summary "
148			echo " All $nAll"
149			echo " Done $nDone"
150			echo " Missing $nMissing"
151			echo ""
152	paus	1.1
153			# construct our job
154	paus	1.11	nFiles=`wc -l $condorOutput/$book/$dataset/fileList.$$.txt \| cut -d ' ' -f1`
155	paus	1.1	if [ "$nFiles" == "" ] \|\| [ "$nFiles" == "0" ]
156			then
157			echo " "
158			echo " No more files to download. EXIT."
159			exit 0
160			elif [ $nFiles -lt $nCopyProcs ]
161			then
162			nCopyProcs=$nFiles
163			fi
164	paus	1.4	# how many files per job?
165	paus	1.1	nFilesPerJob=$(( $nFiles/$nCopyProcs ))
166	paus	1.4	nFilesTmp=$(( $nFilesPerJob*$nCopyProcs ))
167			if [ $nFilesPerJob == 1 ] && [ $nFiles -gt $nCopyProcs ]
168			then
169			nFilesPerJob=2
170			elif [ $nFilesTmp -lt $nFiles ]
171			then
172			nFilesPerJob=$(( $nFilesPerJob+1 ))
173			fi
174
175	paus	1.10	echo " n files to copy: $nFiles"
176			echo " n files/proc : $nFilesPerJob"
177	paus	1.1
178			i=1
179			next=1
180			last=$nFilesPerJob
181
182	paus	1.3	# make sure condor is properly setup for us
183			if ! [ -z $CONDOR_LOCATION ]
184			then
185			unset CONDOR_LOCATION
186			export CONDOR_CONFIG=/usr/local/condor/etc/condor_config
187			fi
188
189	paus	1.12	# stage in the missing files if it is at CERN
190			if [ "`echo $dataDir \| grep /castor/cern.ch`" != "" ]
191			then
192			echo " scp $condorOutput/$book/$dataset/fileList.$$.txt $TICKET_HOLDER@lxplus.cern.ch:"
193			scp $condorOutput/$book/$dataset/fileList.$$.txt $TICKET_HOLDER@lxplus.cern.ch:
194			echo " ssh $TICKET_HOLDER@lxplus.cern.ch ./stageSample.py --dataDir=$dataDir/$book/$dataset --fileList=fileList.$$.txt"
195			ssh $TICKET_HOLDER@lxplus.cern.ch ./stageSample.py --dataDir=$dataDir/$book/$dataset --fileList=fileList.$$.txt
196			echo " ssh $TICKET_HOLDER@lxplus.cern.ch rm fileList.$$.txt"
197			ssh $TICKET_HOLDER@lxplus.cern.ch rm fileList.$$.txt
198			fi
199
200	paus	1.3	# loop over the condor jobs and submit them
201	paus	1.4	while [ $i -le $nCopyProcs ] && [ $last -le $nFiles ]
202	paus	1.1	do
203			if [ $i == $nCopyProcs ]
204			then
205			last=$nFiles
206			fi
207
208			# say what we are going to submit
209	paus	1.11	echo " downloadFiles.sh $dataDir $book $dataset $target $condorOutput $$ $next $last"
210	paus	1.1
211			logFile=`echo download:$book/$dataset/${next}-${last}.txt \| tr '/' '+'`
212			logFile=/tmp/$logFile
213			rm -f $logFile
214
215			# prepare the condor_submit files
216			cat > submit_$$.cmd <<EOF
217			Universe = vanilla
218	paus	1.7	Requirements = ( (Arch == "INTEL") && (Disk >= DiskUsage) && ((Memory * 1024) >= ImageSize) && (HasFileTransfer) )
219	paus	1.12	Notify_user = $TICKET_HOLDER@mit.edu
220	paus	1.1	Notification = Error
221			Executable = $script
222	paus	1.11	Arguments = $dataDir $book $dataset $target $condorOutput $$ $next $last
223	paus	1.1	Rank = Mips
224			GetEnv = True
225			Input = /dev/null
226			Output = $condorOutput/$book/$dataset/${next}-${last}.out
227			Error = $condorOutput/$book/$dataset/${next}-${last}.err
228			Log = $logFile
229			should_transfer_files = YES
230			when_to_transfer_output = ON_EXIT
231	paus	1.3
232			+AccountingGroup = "group_cmsuser.cmsu0284"
233
234	paus	1.1	Queue
235			EOF
236
237	paus	1.3	#+x509userproxysubject = $DN
238
239	paus	1.1	# submit the jobs
240			condor_submit submit_$$.cmd >& /dev/null #>& lastSub
241			#cat submit_$$.cmd
242			rm submit_$$.cmd
243
244			# update counters
245			next=$(( $next + $nFilesPerJob ))
246			last=$(( $last + $nFilesPerJob ))
247			i=$(( $i + 1 ))
248			done
249
250			exit 0