source: trunk/Mars/datacenter/scripts/dbchk@ 13591

#!/bin/sh
#
# ========================================================================
#
# *
# * This file is part of MARS, the MAGIC Analysis and Reconstruction
# * Software. It is distributed to you in the hope that it can be a useful
# * and timesaving tool in analysing Data of imaging Cerenkov telescopes.
# * It is distributed WITHOUT ANY WARRANTY.
# *
# * Permission to use, copy, modify and distribute this software and its
# * documentation for any purpose is hereby granted without fee,
# * provided that the above copyright notice appear in all copies and
# * that both that copyright notice and this permission notice appear
# * in supporting documentation. It is provided "as is" without express
# * or implied warranty.
# *
#
#
#   Author(s): Thomas Bretz  03/2007 <mailto:tbretz@astro.uni-wuerzburg.de>
#   Author(s): Daniela Dorner  02/2008 <mailto:dorner@astro.uni-wuerzburg.de>
#   Author(s): Stefan Ruegamer  02/2008 <mailto:snruegam@astro.uni-wuerzburg.de>
#
#   Copyright: MAGIC Software Development, 2000-2008
#
#
# ========================================================================
#
# This script searches for crashed runs, sequences or datasets in the db
# as well as obsolete or missing sequence files and corresponding callisto
# and star folders. In the third and fourth check the consistency between
# the Run and Sequence DB is checked.
#
#
# TODO
#
# Spped up with SQL: FIELD(25123, 1, 2, 3, 4, 5, ..., 1000200);
# Cross check sequences in Sequence and other databases
# added switch which allows to correct for the problems immediatly
# unify the checks (especially 1b and 1c)
# for the sequences better use 0* instead of 0+ ?


source `dirname $0`/sourcefile
printprocesslog "INFO starting $0"
program=dbchk

getdbsetup
alias mymysql='mysql -s -u $us --password=$pw --host=$ho $db'

# check for crashed nights, runs, sequences and datasets
printprocesslog "INFO Checking if something is crashed on nightly basis" 
nights=`echo "SELECT fDate FROM SequenceBuildStatus WHERE isnull(fFailedTime) and not isnull(fStartTime) and adddate(fStartTime, Interval 12 HOUR) < Now()" | mymysql`
if [ ! "$nights" = "" ]
then
   printprocesslog "WARN For the following nights something seems to be crashed. Please check manually: $nights"
else
   printprocesslog "INFO Nothing found." 
fi

printprocesslog "INFO Checking if something is crashed on run basis" 
cruns=`echo "SELECT fRunNumber FROM RunProcessStatus WHERE isnull(fFailedTime) and not isnull(fStartTime) and adddate(fStartTime, Interval 12 HOUR) < Now()" | mymysql`
if [ ! "$cruns" = "" ]
then
   printprocesslog "WARN The following runs seem to be crashed. Please check manually: $cruns"
else
   printprocesslog "INFO Nothing found." 
fi

printprocesslog "INFO Checking if something is crashed on sequence basis" 
csequences=`echo "SELECT fSequenceFirst FROM SequenceProcessStatus WHERE isnull(fFailedTime) and not isnull(fStartTime) and adddate(fStartTime, Interval 12 HOUR) < Now()" | mymysql`
if [ ! "$csequences" = "" ]
then
   printprocesslog "WARN The following sequences seem to be crashed. Please check manually: $csequences"
else
   printprocesslog "INFO Nothing found." 
fi

printprocesslog "INFO Checking if something is crashed on dataset basis" 
cdatasets=`echo "SELECT fDataSetNumber FROM DataSetProcessStatus WHERE isnull(fFailedTime) and not isnull(fStartTime) and adddate(fStartTime, Interval 12 HOUR) < Now()"  | mymysql`
if [ ! "$cdatasets" = "" ]
then
   printprocesslog "WARN The following datasets seem to be crashed. Please check manually: $cdatasets"
else
   printprocesslog "INFO Nothing found." 
fi

# CHECK 1
printprocesslog "INFO Checking if all sequence files have a corresponding entry in Sequences" 
files=`find $sequpath -type f`
for file in $files
do
   sequence=`echo $file | sed -e "s/^.*\/sequence0\+\([0-9]\+\)\.txt$/\1/"`
   if [ "$sequence" = "" ] || [ "$sequence" = "$file" ]
   then
      printprocesslog "ERROR No sequence file: $file" 
      continue
   fi
 
   var=`echo SELECT fSequenceFirst FROM Sequences WHERE fSequenceFirst=$sequence | mymysql`
   if ! [ "$sequence" = "$var" ]
   then
      printprocesslog "ERROR Sequence-File $sequence exists, but it is not in Sequences (DB)." 
      continue
   fi
done

# CHECK 1b (callisto)
printprocesslog "INFO Checking if all sequences in ${datapath}/callisto have a corresponding sequence in Sequence" 
dirs=`find ${datapath}/callisto -mindepth 2 -maxdepth 2 -type d`
for dir in $dirs
do
   sequence=`echo $dir | sed -e "s/^.*\/0\+\([0-9]\+\)$/\1/"`
   if [ "$sequence" = "" ] || [ "$sequence" = "$dir" ]
   then
      printprocesslog "ERROR Invalid directory: $dir" 
      continue
   fi
   
   var=`echo SELECT fSequenceFirst FROM Sequences WHERE fSequenceFirst=$sequence | mymysql`
   if ! [ "$sequence" = "$var" ]
   then
      printprocesslog "ERROR $dir exists, but no corresponding sequence in Sequences (DB)." 
      continue
   fi
done

# CHECK 1c (star)
printprocesslog "INFO Checking if all sequences in ${datapath}/star have a corresponding sequence in Sequence" 
dirs=`find ${datapath}/star -mindepth 2 -type d`
for dir in $dirs
do
   sequence=`echo $dir | sed -e "s/^.*\/0\+\([0-9]\+\)$/\1/"`
   if [ "$sequence" = "" ] || [ "$sequence" = "$dir" ]
   then
      printprocesslog "ERROR Invalid directory: $dir" 
      continue
   fi
   
   var=`echo SELECT fSequenceFirst FROM Sequences WHERE fSequenceFirst=$sequence | mymysql`
   if ! [ "$sequence" = "$var" ]
   then
      printprocesslog "ERROR $dir exists, but no corresponding sequence in Sequences (DB)." 
      continue
   fi
done

# CHECK 2
printprocesslog "INFO Checking if all sequences in Sequences have a corresponding sequence files" 
sequences=`echo SELECT fSequenceFirst FROM Sequences left join SequenceProcessStatus using (fSequenceFirst,fTelescopeNumber) where not isnull(fSequenceFileWritten) | mymysql`
for sequence in $sequences
do
   var=`find $sequpath -type f -regex .*/sequence0+${sequence}\.txt$`
   if [ "$var" = "" ]
   then
      printprocesslog "ERROR Sequence-File for $sequence not found but in DB." 
   fi
done

# CHECK 3
printprocesslog "INFO Checking if all sequences from Sequences exist RunData" 
sequences=`echo SELECT fSequenceFirst FROM Sequences GROUP BY fSequenceFirst | mymysql`
for sequence in $sequences
do
   res=`echo SELECT fSequenceFirst FROM RunData WHERE fSequenceFirst=$sequence GROUP BY fSequenceFirst | mymysql`
   if ! [ "$sequence" = "$res" ]
   then
      printprocesslog "ERROR Sequence $sequence exists in Sequences but not in RunData (DB)."
      continue
   fi
done

# CHECK 4
printprocesslog "INFO Checking if all sequences from RunData exist in Sequences" 
sequences=`echo SELECT fSequenceFirst FROM RunData WHERE not fSequenceFirst=0 GROUP BY fSequenceFirst | mymysql`
for sequence in $sequences
do
   var=`echo SELECT fSequenceFirst FROM Sequences WHERE fSequenceFirst=$sequence | mymysql`
   if ! [ "$sequence" = "$var" ]
   then
      printprocesslog "ERROR Sequence $sequence exists in RunData but not in Sequences (DB)." 
      continue
   fi
done

# CHECK 5
printprocesslog "INFO Checking if all sequences from RunData exist in SequenceProcessStatus" 
sequences=`echo SELECT fSequenceFirst FROM RunData WHERE not fSequenceFirst=0 GROUP BY fSequenceFirst | mymysql`
for sequence in $sequences
do
   var=`echo SELECT fSequenceFirst FROM SequenceProcessStatus WHERE fSequenceFirst=$sequence | mymysql`
   if ! [ "$sequence" = "$var" ]
   then
      printprocesslog "ERROR Sequence $sequence exists in RunData but not in SequenceProcessStatus (DB)." 
      continue
   fi
done

finish