Esempio n. 1
0
 def doJson(self,lfn=None, jobs=None,type='input',ds=None,table='new',limit=1000,jobstatus=None,site=None,jobtype='production',days=1,user=None,select=None):
    """ 
       Show  the list of Panda id with LFN or versa verse LFNs by Panda IDs
       <ul>
       <li><code>lfn</code>   - the list of the comma separated files 
       <li><code>ds</code>    - the list of the comma separated datasets 
       <li><code>jobs</code>  - the list of the comma separated Panda's job IDs
       <li><code>table</code>  = "new" (default) look up the records for last 3 days
                      <br> ="old" - look up the records those more than 3 days old (slow)
                      <br> ="deep" - look up the "old" and "new" tables (slow)
       <li><code>type</code>  - the type selector. <br>
                             =  'input - the default value<br>
                             =  '*' | 'all' - list all types available.
        <li><code>jobstatus</code> = the comma separated list of the job status to filter
        <br>For example: 'defined, waiting,assigned,activated,sent,starting,running'
        <li><code>site</code> = the comma separated list of the sizte to list the jobs from         
        <br> For example 'UTA_SWT2'
        <li><code>jobtype</code> = the comma separated list of the job type to filter
        <br> For example, "analysis, production"
        <li><code>days</code> = the number of the days to look up the list of the jobs if either 'jobstatus' or 'site'  parameter is defined
        <li><code>user</code> = the comma separated list of the usernames . 
        <br>NB. The names may contain the the wildcard symbol '*'. Be aware the wildcard slows the search down
       </ul>
    """ 
    title = 'The list of files for the  '
    if jobstatus and jobstatus.strip()   =='': jobstatus = None
    if site and site.strip()   =='':        site = None
    if lfn and lfn.strip()   =='': lfn = None
    if jobs and isinstance(jobs,str) and jobs.strip() =='': jobs = None
    if ds and ds.strip()     =='': ds=None
    if type and type.strip() =='': type='all'
    if  lfn==None and jobs==None and ds==None and jobstatus==None and site==None:
         self.publishTitle("Ambigios query: lfn=%(lfn)s; pandaid=%(pandaid)s either lfn or padaid can be defined. One can not define lfn and pandaid at once" % { 'lfn': lfn, 'pandaid' : jobs} )
         self.publish('<h2>Check the input parameters. Click the "?" to see the API documentaion<h2>', role=pmRoles.html())
    else:
       nav = ''
       if limit:
           nav += "Limit %s rows." % limit
       if type=='*' or type==None: type = 'all'
       if lfn != None:
          self.publishTitle("The list of the PANDA jobs with the LFN of the '%s' type provided" % type)
          if not '*' in lfn:  # disregard the jobtype parameter
             if utils.isFilled(jobtype): 
                nav += " Disregarding the jobtype='%s' default parameter" % jobtype 
                jobtype = None
       if ds != None:   
          self.publishTitle("The list of the PANDA jobs with the DATASETs of the '%s' type provided" % type)
       if jobs!=None:
          self.publishTitle("The list of the '%s' LFN with the PANDA Job IDs provided" % type)
       if utils.isFilled(nav):
          self.publishNav(nav)
       main = {}
       main["buffer"] = {}
       main["buffer"]["method"] = "joblfn"
       main["buffer"]["params"] = (lfn if lfn!=None else '',jobs if jobs!= None else '' ) 
       if jobs != None: main["buffer"]["jobs"] = jobs
       main["buffer"]["type"] = False
       if (utils.isFilled(jobstatus) or utils.isFilled(site) or utils.isFilled(user)) and  not utils.isFilled(jobs):
           tables = ['atlas_panda.jobsArchived4','atlas_panda.jobsActive4','atlas_panda.jobsWaiting4','atlas_panda.jobsDefined4']
           r = pmt.getJobIds(site, jobtype,jobstatus,table=tables,days=days,username=user)
           jobs = [i[0] for i in r['rows']]
       if not utils.isFilled(select):
          select = []    
          if jobs == None or ( not isinstance(jobs,int) and len(jobs) > 1): select.append('pandaid');
          select += ['type', 'lfn', 'fsize', 'dataset', 'guid', 'scope', 'destinationse']
       else:
          select = utils.parseArray(select);         
       main["buffer"]["data"] = pmt.getJobLFN(select=','.join(select),table=table,lfn=lfn,pandaid=jobs,type=type,ds=ds,limit=limit)
       self.publish(main)
       self.publish( "%s/%s" % (self.server().fileScriptURL(),"taskBuffer/%s.js" % "getJobLFN"),role=pmRoles.script())
Esempio n. 2
0
 def doJson(self,table=None,column=None,db='pmt',doc=None):
    """ 
        Describe the selected Panda table / column <br>
        <code>table  = the Db table name regexp comma-separated patterns to check</code><br>
        <code>column = the Db table column name regexp comma-separated patterns to find</code><br>
        <code>doc    = '*|all' - document all columns selected if available<br>
             = 'comma-separated list of the regexp' is to display the columns with  the comments matched the regexp provided
        </code><br>
    """ 
    title = 'Describe the Panda Db '
    inp = False
    if table != None  and table != 'None': 
       title += "'%s*'" % table
       inp = True
       title +=" table"
    else:   
       title +=" tables"
    if column != None and column !='None': 
        title += " with '%s*' column " % column
        inp = True
    main = {}
    if  inp:      
       self.publishTitle(title)
       timer = Stopwatch.Stopwatch()
       main["buffer"] = {}
       main["buffer"]["db"] = db
       dbAccess = self._db.get(db)
       if dbAccess != None:
          main["buffer"]["method"] = "describe"
          main["buffer"]["params"] = (table,column,doc) 
          main["buffer"]["type"] = False
          if not utils.isFilled(doc):
             tdes = dbAccess.describe(table,column)
             iclmn = utils.name2Index(tdes['header'],'COLUMN_NAME')
             rows = tdes['rows']
             for h in rows:
                h[iclmn] = nmap.get(h[iclmn])
             main["buffer"]["data"] = tdes
          else:
             docrows = []
             if utils.isFilled(table):
                tdoc = dbAccess.pandatabs(table,comments=doc,select="table_name,'Table Description' as COLUMN_NAME,comments")
                iclmn = utils.name2Index(tdoc['header'],'COLUMN_NAME')
                docrows = tdoc['rows']
                for h in docrows:
                   h[iclmn] = nmap.get(h[iclmn])
             tcol = dbAccess.pandacols(column,table,select='TABLE_NAME,COLUMN_NAME,COMMENTS', comments=doc)
             iclmn = utils.name2Index(tcol['header'],'COLUMN_NAME')
             for h in tcol['rows']:
                h[iclmn] = nmap.get(h[iclmn])
             tcol['rows'] += docrows
             main["buffer"]["data"] = tcol
             self.publish( {'s-maxage':60000,'max-age': 6000}, role=pmRoles.cache())
       else:
          self.publishTitle("Error: 'describe' -  unknown Db name '%s' has been provided" % db)
          self.publish( {'s-maxage':0,'max-age': 0}, role=pmRoles.cache())
    else:
       self.publishTitle("Error: 'describe' -  no parameter has been provided. Define one at least. It should be either 'column' or 'table'")
       self.publish( {'s-maxage':0,'max-age': 0}, role=pmRoles.cache())
    self.publish(main)
    self.publish( "%s/%s" % (self.server().fileScriptURL(),"taskBuffer/%s.js" % "describe"),role=pmRoles.script() )
Esempio n. 3
0
    def doJobs(
        self,
        pandaid=None,
        params=None,
        jobparam=None,
        table=None,
        hours=None,
        tstart=None,
        tend=None,
        field="modificationTime",
        days=None,
        format=None,
        jobtype=None,
        testsite=None,
        limit=1500,
        summary=None,
        minjob=1,
        dump=None,
        region=None,
        lfn=None,
        showcpu=None,
    ):
        ## Perform the query
        main = {}
        errorcolcheck = False
        selection = "*"
        vals = self.extraValuesFilled(self._alias)

        title = "PanDA Job(s)"
        if utils.isFilled(pandaid):
            hours = None
            days = None
            title += " %s " % pandaid
        if days != None:
            title += " for %.2f Day" % float(days)
            if days > 1:
                title += "s"
        if hours != None:
            if days == None:
                title += " for "
            title += " %d Hour" % int(hours)
            if hours > 1:
                title += "s"

        if vals != None:
            for k in vals.keys():
                if vals[k] == "undefined":
                    vals[k] = None
                elif (
                    vals[k] != None
                    and k.lower() == "jobstatus"
                    and (
                        vals[k].lower().find("prerun") >= 0
                        or vals[k].lower().find("finishing") >= 0
                        or vals[k].lower().find("unassigned") >= 0
                        or vals[k].lower().find("retried") >= 0
                    )
                ):
                    stvals = utils.parseArray(vals[k])
                    statvals = []
                    for v in stvals:
                        if v.lower() == "prerun":
                            statvals += ["defined", "assigned", "waiting", "activated", "pending", "sent", "starting"]
                        elif v.lower() == "finishing":
                            statvals += ["holding", "transferring"]
                        elif v.lower() == "unassigned":
                            vals["computingSite"] = "NULL"
                            statvals += ["cancelled"]
                        elif v.lower() == "retried":
                            # See the slide 3 from Tadashi: https://indico.cern.ch/getFile.py/access?contribId=13&sessionId=11&resId=0&materialId=slides&confId=119171
                            vals["taskBufferErrorCode"] = [106, 107]
                            statvals += ["failed"]
                        else:
                            statvals += [v]
                        vals[k] = ",".join(statvals)
            cleanVal = dict((k, v) for k, v in vals.iteritems() if v is not None)
            vals = cleanVal
        if vals == None:
            vals = {}
        nav = ""
        certificate = None
        certusr = None
        try:
            cert = self.server().certificate()
            if cert != None:
                certusr = pmt.cleanUserID(cert)
                main["ceruser"] = certusr
        except:
            pass
        if vals.get("prodUserName") == "auto" and certusr != None:
            if cert != None:
                vals["prodUserName"] = certusr
        for v in vals:
            nav += "<b>%s</b>=%s " % (v, vals[v])
        # nav += "<span id=linkjobid  stule='display:inline'></span>"
        # nav +="<script>$(document).ready(function(){$('#linkjobid').html(utils().linkJobs(s.cloud,jobtype"
        # nav += jobtype if jobtype!=None else "undefined"
        # nav += ", s.jobStatus, "
        # nav += "%d" % hours if hours!=None else "undefined"
        # nav += ", s.computingSite,'Classic Page' ))});</script>";
        if not utils.isFilled(summary):
            nav += "<b>%s</b>=%s " % ("limit", limit)

        self.publishNav(nav)
        if len(vals) == 0:
            vals = None
        if vals == None and jobparam == None:
            main["Description"] = self.description()
        else:
            if lfn != None and lfn != "undefined":
                lfnselect = "pandaid,lfn"
                lfnjobs = pmt.getJobLFN(select=lfnselect, table="new", lfn=lfn, pandaid=pandaid, type="*", limit=limit)
                if lfnjobs["rows"] > 0:
                    pandaid = [ljobs[0] for ljobs in lfnjobs["rows"]]
            selection = {}
            if vals != None:
                selection.update(vals)
            if pandaid != None:
                selection.update({nmap.get("pandaid"): pandaid})
            if region != None:
                selection.update({"region": region})
            main["Selection"] = selection
            conditions = vals
            if conditions == None:
                conditions = {}
            cs = conditions.get("computingSite")
            if cs == None:
                site4region = self.getSite4Region(region)
                if site4region != None and len(site4region) > 0:
                    conditions["computingSite"] = site4region
            else:
                pass  # ignore the region parameter
            if utils.isValid(summary):
                jobtype = (
                    "production"
                    if jobtype == None
                    and jobparam.lower().find("taskid") >= 0
                    and jobparam.lower().find("jeditaskid") < 0
                    else jobtype
                )
                jobs = pmt.getJobsAtt(
                    pandaid,
                    jobparam,
                    conditions,
                    None,
                    hours,
                    tstart,
                    tend,
                    field,
                    days,
                    format,
                    jobtype,
                    testsite,
                    None,
                    True,
                )
                rows = jobs["rows"]
                indx = []
                if len(rows) > 30:
                    for i, r in enumerate(rows):
                        try:
                            if r[1] > minjob:
                                indx.append(i)
                                continue
                        except:
                            pass
                    cleanrows = [rows[i] for i in indx]
                    jobs["rows"] = cleanrows
                istatus = utils.name2Index(jobs["header"], "jobstatus")
                if istatus != None:
                    statOrder = utils.name2Index(utils.jobStates)

                    def byStatus(a, b):
                        return statOrder[a[istatus]] - statOrder[b[istatus]]

                    jobs["rows"] = sorted(jobs["rows"], cmp=byStatus)
            else:
                if jobparam == self._allparams and conditions.get(self._jobParametersName) == None:
                    conditions[self._jobParametersName] = "any"
                jobs = pmt.getJobsAtt(
                    pandaid,
                    "PandaID," + jobparam,
                    conditions,
                    None,
                    hours,
                    tstart,
                    tend,
                    field,
                    days,
                    format,
                    jobtype,
                    testsite,
                    limit,
                )
                errorcolcheck = True
            if len(jobs) > 0:
                header = [nmap.get(h) for h in jobs["header"]]
                info = jobs["rows"]
                main["header"] = header  #  + self._extraFields
                timecols = []
                errdict = None
                errorcol = not errorcolcheck
                for i, h in enumerate(header):
                    if "time" in h.lower() and "cpuconsumptiontime" != h.lower():
                        timecols.append(i)
                    if not errorcol and "error" in h.lower():
                        errorcol = True
                ##               # add an artificial retryID
                ##               header += ['retryID']
                headindx = utils.name2Index(header)
                creationtDict = {}
                histograms = {}
                prodSourceLabel = headindx.get("prodSourceLabel")
                jobStatus = headindx.get("jobStatus")
                status4title = conditions.get("jobStatus")
                if not utils.isFilled(status4title):
                    if len(info) == 1 and jobStatus != None:
                        status4title = info[0][jobStatus]
                if utils.isFilled(status4title):
                    title = status4title[0].upper() + status4title[1:] + " " + title
                prodtype = "production" in jobtype.lower() if jobtype != None else True
                taskidsum = prodtype
                if not utils.isValid(summary) and dump != "yes":
                    jsets = setutils.factory(header, info, self._errorCodes, self._siteId, self._siteNicks)
                    main["jobset"] = jsets.jobsets()
                for r in info:
                    if not taskidsum and ((prodSourceLabel and r[prodSourceLabel] == "managed") and prodtype):
                        taskidsum = True
                    if errorcol and errorcolcheck:
                        errdict = self.getErrors(r, headindx, errdict)
                    for t in timecols:
                        r[t] = utils.epochTime(r[t])
                ##                  r += [None] # retryID
                # row = self.jobsetTime(utils.zip(header(info[r])),creationtDict,histograms)
                # info[r] = self.addFields(utils.zip(header(info[r])),creationtDict,histograms)
                # Action            cleanUserID SSL_CLIENT_S_DN
                main["info"] = info
                if errorcolcheck:
                    main["errorCodes"] = errdict
                main["colNames"] = sorted(header)
                main["jobsummary"] = copy.copy(self._jobSum)
                if taskidsum:
                    main["jobsummary"] += [["Task ID", "taskID"]]
                #               print utils.lineInfo(), "===============", self._description.get("JEDITASKID"), self._description.get("JediTaskID")
                #               if self._description.get("JEDITASKID") or self._description.get("JediTaskID"):
                #                  main['jobsummary'] += [['Jedi ID','JediTaskID']]
                elif jobtype != "jedi":
                    main["jobsummary"] += [["Jobsets", "jobsetID"]]  #  need constrain
                elif jobtype == "analysis":
                    main["jobsummary"] += [["Analysis Tasks", "destinationDBlock"]]
                if showcpu == "yes":
                    main["jobsummary"] += [["CPU Type", "cpuConsumptionUnit"]]
        self.publishTitle(title)
        return main
Esempio n. 4
0
 def doJson(
     self,
     job=None,
     taskid=None,
     tstart=None,
     tend=None,
     hours=71,
     days=None,
     jobparam=None,
     limit=1500,
     summary=None,
     minjob=1,
     jobtype=None,
     timeField="modificationTime",
     plot="no",
     dump=None,
     tasktype=None,
     status=None,
     region=None,
     lfn=None,
     showcpu=None,
 ):
     """Display the job information: 
  <p>Sample queries:
  <ul>
   <li><a href='http://pandamon.cern.ch/jobinfo?tstart=2013-04-07 00:00&amp;tend=2013-04-07 00:03&amp;hours='>http://pandamon.cern.ch/jobinfo/?tstart=2011-06-07 00:00&amp;tend=2011-06-07 00:01</a>
    <li><a href='http://pandamon.cern.ch/jobinfo?tstart=2013-04-07 00:00&amp;tend=2013-04-07 00:03&amp;hours=&amp;processingType=simul'>http://pandamon.cern.ch/jobinfo/?tstart=2011-06-07 00:00&amp;tend=2011-06-07 00:01&[email protected]</a>
    <li><a href='http://pandamon.cern.ch/jobinfo?job=1246240909'>http://pandamon.cern.ch/jobinfo?job=1246240909</a>
   </ul>
       <hr>
       Supported query parameters are as follows:
       <table>
       <tr><td>taskid</td><td> - jobset / task id</td></tr>
       <tr><td>job</td><td> - comma separated list of jobs' PandaID</td></tr>
       <tr><td>jobtype</td><td> - select the 'jobtype' jobs: analysys, production, install, test </td></tr>
       <tr><td>tstart=None</td><td>  - start time </td></tr>
       <tr><td>tend=None</td><td>    - end time</td></tr>
       <tr><td>hours=N </td><td> - last (N hours + N days) or N hours after 'tstart'  or N hours before 'tend' </td></tr> 
       <tr><td>days=N</td><td>   - last (N hours + N days)  (see the hours parameter documentation</td></tr>
       <tr><td>lfn=</td><td>   -  Select the PanDA jobs using the files from the comma-separated lfn list </td></tr>
       <tr><td>jobparam=</td><td>The comma separated list of the <a title='Click to see the full list of the jobinfo parameters available' href='http://pandamon.cern.ch/describe?table=^JOBSARCHIVED4&doc=*'>job parameters</a>.<br> The  default should satisfy the interactive users.<br>
       One needs to customize jobparam to build his/her <a href='https://twiki.cern.ch/twiki/bin/view/PanDA/PandaPlatform#Custom_Monitoring_Client_Code'>own client</a></td></tr>
       <tr><td>limit=</d><td>   - max number of the Db records to fetch from  each job Db table. Since there are 4 tables the real number of the rows can be 4 times as many as that limit</td></tr>
       <tr><td><a title='Click to see the full list of the jobinfo parameters available' href='http://pandamon.cern.ch/describe?table=^JOBSARCHIVED4&doc=*'>PARAM</a>=value</td><td>-  comma-separated list of the value <a href ='https://twiki.cern.ch/twiki/bin/view/PanDA/PandaPlatform#API_parameter_value_format'> patterns</a> to filter the job records.<br>
       For example: <ul><li><a href='http://pandamon.cern.ch/jobinfo?processingType=simul'>http://pandamon.cern.ch/jobinfo?processingType=simul</a>
       <li> <a href='http://pandamon.cern.ch/jobinfo?limit=2&hours=4&processingType=ganga*,test'>http://pandamon.cern.ch/jobinfo?limit=2&hours=4&processingType=ganga*,test</a></ul>
       <tr><td>summary</td><td>Create the summary for the first member of the jobparam list<br>
       For example: <a href='http://pandamon.cern.ch/jobinfo?processingType=simul&jobparam=jobStatus&summary=yes'>http://pandamon.cern.ch/jobinfo?processingType=simul&jobparam=jobStatus&summary=yes</a>
       </td></tr>
       <tr><td>dump</td><td> - 'yes' - produce the flat list of job info  records suitable for CLI applications</td></tr>
       <tr><td>region</td><td> - the comma separated list of the regions </td></tr>
       <tr><td>showcpu</td><td> - 'yes' to show the CPU type used </td></tr>
       </table>
   """
     #       % { 'url' : self.server().branchUrl() }
     if job == "undefined":
         job = None
     if limit == "undefined":
         limit = None
     if summary == "undefined":
         summary = None
     if jobparam == "undefined":
         jobparam = None
     if tstart == "undefined":
         tstart = None
     if tend == "undefined":
         tend = None
     if tasktype == "prod":
         jobtype = "production"
     if jobtype == "null":
         jobtype = None
     if jobparam == None:
         if job != None:
             jobparam = self._allparams
         else:
             jobparam = ",".join(self._colNames)
     if timeField == None or timeField == "undefined" or timeField == "":
         timeField = "modificationTime"
     if self.isValid(jobparam):
         self.publish(
             self.doJobs(
                 job,
                 tstart=tstart,
                 tend=tend,
                 hours=hours,
                 days=days,
                 jobparam=jobparam,
                 limit=limit,
                 field=timeField,
                 jobtype=jobtype,
                 summary=summary,
                 minjob=minjob,
                 dump=dump,
                 region=region,
                 lfn=lfn,
                 showcpu=showcpu,
             )
         )
         self.publish("%s/%s" % (self.server().fileScriptURL(), "jobInfoRender.js"), role=pmRoles.script())
         self.publish({"s-maxage": 600, "max-age": 600}, role=pmRoles.cache())
     else:
         title = "Job data "
         if utils.isFilled(job):
             hours = None
             days = None
             title += " %s " % job
         if days != None:
             title += " for %.2f day" % float(days)
             if days > 1:
                 title += "s"
         if hours != None:
             if days == None:
                 title += " for "
             title += " %d hour" % int(hours)
             if hours > 1:
                 title += "s"
         self.publishTitle(title)
Esempio n. 5
0
 def doJson(self, table=None, column=None, db='pmt', doc=None):
     """ 
       Describe the selected Panda table / column <br>
       <code>table  = the Db table name regexp comma-separated patterns to check</code><br>
       <code>column = the Db table column name regexp comma-separated patterns to find</code><br>
       <code>doc    = '*|all' - document all columns selected if available<br>
            = 'comma-separated list of the regexp' is to display the columns with  the comments matched the regexp provided
       </code><br>
   """
     title = 'Describe the Panda Db '
     inp = False
     if table != None and table != 'None':
         title += "'%s*'" % table
         inp = True
         title += " table"
     else:
         title += " tables"
     if column != None and column != 'None':
         title += " with '%s*' column " % column
         inp = True
     main = {}
     if inp:
         self.publishTitle(title)
         timer = Stopwatch.Stopwatch()
         main["buffer"] = {}
         main["buffer"]["db"] = db
         dbAccess = self._db.get(db)
         if dbAccess != None:
             main["buffer"]["method"] = "describe"
             main["buffer"]["params"] = (table, column, doc)
             main["buffer"]["type"] = False
             if not utils.isFilled(doc):
                 tdes = dbAccess.describe(table, column)
                 iclmn = utils.name2Index(tdes['header'], 'COLUMN_NAME')
                 rows = tdes['rows']
                 for h in rows:
                     h[iclmn] = nmap.get(h[iclmn])
                 main["buffer"]["data"] = tdes
             else:
                 docrows = []
                 if utils.isFilled(table):
                     tdoc = dbAccess.pandatabs(
                         table,
                         comments=doc,
                         select=
                         "table_name,'Table Description' as COLUMN_NAME,comments"
                     )
                     iclmn = utils.name2Index(tdoc['header'], 'COLUMN_NAME')
                     docrows = tdoc['rows']
                     for h in docrows:
                         h[iclmn] = nmap.get(h[iclmn])
                 tcol = dbAccess.pandacols(
                     column,
                     table,
                     select='TABLE_NAME,COLUMN_NAME,COMMENTS',
                     comments=doc)
                 iclmn = utils.name2Index(tcol['header'], 'COLUMN_NAME')
                 for h in tcol['rows']:
                     h[iclmn] = nmap.get(h[iclmn])
                 tcol['rows'] += docrows
                 main["buffer"]["data"] = tcol
                 self.publish({
                     's-maxage': 60000,
                     'max-age': 6000
                 },
                              role=pmRoles.cache())
         else:
             self.publishTitle(
                 "Error: 'describe' -  unknown Db name '%s' has been provided"
                 % db)
             self.publish({
                 's-maxage': 0,
                 'max-age': 0
             },
                          role=pmRoles.cache())
     else:
         self.publishTitle(
             "Error: 'describe' -  no parameter has been provided. Define one at least. It should be either 'column' or 'table'"
         )
         self.publish({'s-maxage': 0, 'max-age': 0}, role=pmRoles.cache())
     self.publish(main)
     self.publish(
         "%s/%s" %
         (self.server().fileScriptURL(), "taskBuffer/%s.js" % "describe"),
         role=pmRoles.script())
Esempio n. 6
0
    def doJobs(self,
               pandaid=None,
               params=None,
               jobparam=None,
               table=None,
               hours=None,
               tstart=None,
               tend=None,
               field='modificationTime',
               days=None,
               format=None,
               jobtype=None,
               testsite=None,
               limit=1500,
               summary=None,
               minjob=1,
               dump=None,
               region=None,
               lfn=None,
               showcpu=None):
        ## Perform the query
        main = {}
        errorcolcheck = False
        selection = '*'
        vals = self.extraValuesFilled(self._alias)

        title = 'PanDA Job(s)'
        if utils.isFilled(pandaid):
            hours = None
            days = None
            title += " %s " % pandaid
        if days != None:
            title += ' for %.2f Day' % float(days)
            if days > 1: title += "s"
        if hours != None:
            if days == None: title += " for "
            title += ' %d Hour' % int(hours)
            if hours > 1: title += "s"

        if vals != None:
            for k in vals.keys():
                if vals[k] == 'undefined': vals[k] = None
                elif vals[k] != None and k.lower() == 'jobstatus' and (
                        vals[k].lower().find('prerun') >= 0
                        or vals[k].lower().find('finishing') >= 0
                        or vals[k].lower().find('unassigned') >= 0
                        or vals[k].lower().find('retried') >= 0):
                    stvals = utils.parseArray(vals[k])
                    statvals = []
                    for v in stvals:
                        if v.lower() == 'prerun':
                            statvals += [
                                'defined', 'assigned', 'waiting', 'activated',
                                'pending', 'sent', 'starting'
                            ]
                        elif v.lower() == 'finishing':
                            statvals += ['holding', 'transferring']
                        elif v.lower() == 'unassigned':
                            vals['computingSite'] = 'NULL'
                            statvals += ['cancelled']
                        elif v.lower() == 'retried':
                            # See the slide 3 from Tadashi: https://indico.cern.ch/getFile.py/access?contribId=13&sessionId=11&resId=0&materialId=slides&confId=119171
                            vals['taskBufferErrorCode'] = [106, 107]
                            statvals += ['failed']
                        else:
                            statvals += [v]
                        vals[k] = ','.join(statvals)
            cleanVal = dict(
                (k, v) for k, v in vals.iteritems() if v is not None)
            vals = cleanVal
        if vals == None: vals = {}
        nav = ''
        certificate = None
        certusr = None
        try:
            cert = self.server().certificate()
            if cert != None:
                certusr = pmt.cleanUserID(cert)
                main['ceruser'] = certusr
        except:
            pass
        if vals.get('prodUserName') == 'auto' and certusr != None:
            if cert != None: vals['prodUserName'] = certusr
        for v in vals:
            nav += "<b>%s</b>=%s " % (v, vals[v])
        # nav += "<span id=linkjobid  stule='display:inline'></span>"
        # nav +="<script>$(document).ready(function(){$('#linkjobid').html(utils().linkJobs(s.cloud,jobtype"
        # nav += jobtype if jobtype!=None else "undefined"
        # nav += ", s.jobStatus, "
        # nav += "%d" % hours if hours!=None else "undefined"
        # nav += ", s.computingSite,'Classic Page' ))});</script>";
        if not utils.isFilled(summary):
            nav += "<b>%s</b>=%s " % ('limit', limit)

        self.publishNav(nav)
        if len(vals) == 0: vals = None
        if vals == None and jobparam == None:
            main['Description'] = self.description()
        else:
            if lfn != None and lfn != "undefined":
                lfnselect = 'pandaid,lfn'
                lfnjobs = pmt.getJobLFN(select=lfnselect,
                                        table='new',
                                        lfn=lfn,
                                        pandaid=pandaid,
                                        type='*',
                                        limit=limit)
                if lfnjobs['rows'] > 0:
                    pandaid = [ljobs[0] for ljobs in lfnjobs['rows']]
            selection = {}
            if vals != None: selection.update(vals)
            if pandaid != None:
                selection.update({nmap.get('pandaid'): pandaid})
            if region != None: selection.update({'region': region})
            main['Selection'] = selection
            conditions = vals
            if conditions == None: conditions = {}
            cs = conditions.get('computingSite')
            if cs == None:
                site4region = self.getSite4Region(region)
                if site4region != None and len(site4region) > 0:
                    conditions['computingSite'] = site4region
            else:
                pass  # ignore the region parameter
            if utils.isValid(summary):
                jobtype = 'production' if jobtype == None and jobparam.lower(
                ).find('taskid') >= 0 and jobparam.lower().find(
                    'jeditaskid') < 0 else jobtype
                jobs = pmt.getJobsAtt(pandaid, jobparam, conditions, None,
                                      hours, tstart, tend, field, days, format,
                                      jobtype, testsite, None, True)
                rows = jobs['rows']
                indx = []
                if len(rows) > 30:
                    for i, r in enumerate(rows):
                        try:
                            if r[1] > minjob:
                                indx.append(i)
                                continue
                        except:
                            pass
                    cleanrows = [rows[i] for i in indx]
                    jobs['rows'] = cleanrows
                istatus = utils.name2Index(jobs['header'], 'jobstatus')
                if istatus != None:
                    statOrder = utils.name2Index(utils.jobStates)

                    def byStatus(a, b):
                        return statOrder[a[istatus]] - statOrder[b[istatus]]

                    jobs['rows'] = sorted(jobs['rows'], cmp=byStatus)
            else:
                if jobparam == self._allparams and conditions.get(
                        self._jobParametersName) == None:
                    conditions[self._jobParametersName] = 'any'
                jobs = pmt.getJobsAtt(pandaid, 'PandaID,' + jobparam,
                                      conditions, None, hours, tstart, tend,
                                      field, days, format, jobtype, testsite,
                                      limit)
                errorcolcheck = True
            if len(jobs) > 0:
                header = [nmap.get(h) for h in jobs['header']]
                info = jobs['rows']
                main['header'] = header  #  + self._extraFields
                timecols = []
                errdict = None
                errorcol = not errorcolcheck
                for i, h in enumerate(header):
                    if 'time' in h.lower(
                    ) and 'cpuconsumptiontime' != h.lower():
                        timecols.append(i)
                    if not errorcol and 'error' in h.lower():
                        errorcol = True
##               # add an artificial retryID
##               header += ['retryID']
                headindx = utils.name2Index(header)
                creationtDict = {}
                histograms = {}
                prodSourceLabel = headindx.get('prodSourceLabel')
                jobStatus = headindx.get('jobStatus')
                status4title = conditions.get('jobStatus')
                if not utils.isFilled(status4title):
                    if len(info) == 1 and jobStatus != None:
                        status4title = info[0][jobStatus]
                if utils.isFilled(status4title):
                    title = status4title[0].upper(
                    ) + status4title[1:] + " " + title
                prodtype = 'production' in jobtype.lower(
                ) if jobtype != None else True
                taskidsum = prodtype
                if not utils.isValid(summary) and dump != 'yes':
                    jsets = setutils.factory(header, info, self._errorCodes,
                                             self._siteId, self._siteNicks)
                    main['jobset'] = jsets.jobsets()
                for r in info:
                    if not taskidsum and (
                        (prodSourceLabel and r[prodSourceLabel] == 'managed')
                            and prodtype):
                        taskidsum = True
                    if errorcol and errorcolcheck:
                        errdict = self.getErrors(r, headindx, errdict)
                    for t in timecols:
                        r[t] = utils.epochTime(r[t])
##                  r += [None] # retryID
# row = self.jobsetTime(utils.zip(header(info[r])),creationtDict,histograms)
# info[r] = self.addFields(utils.zip(header(info[r])),creationtDict,histograms)
# Action            cleanUserID SSL_CLIENT_S_DN
                main['info'] = info
                if errorcolcheck:
                    main['errorCodes'] = errdict
                main['colNames'] = sorted(header)
                main['jobsummary'] = copy.copy(self._jobSum)
                if taskidsum:
                    main['jobsummary'] += [['Task ID', 'taskID']]


#               print utils.lineInfo(), "===============", self._description.get("JEDITASKID"), self._description.get("JediTaskID")
#               if self._description.get("JEDITASKID") or self._description.get("JediTaskID"):
#                  main['jobsummary'] += [['Jedi ID','JediTaskID']]
                elif jobtype != 'jedi':
                    main['jobsummary'] += [['Jobsets',
                                            'jobsetID']]  #  need constrain
                elif jobtype == 'analysis':
                    main['jobsummary'] += [[
                        'Analysis Tasks', 'destinationDBlock'
                    ]]
                if showcpu == 'yes':
                    main['jobsummary'] += [['CPU Type', 'cpuConsumptionUnit']]
        self.publishTitle(title)
        return main
Esempio n. 7
0
 def doJson(self,
            job=None,
            taskid=None,
            tstart=None,
            tend=None,
            hours=71,
            days=None,
            jobparam=None,
            limit=1500,
            summary=None,
            minjob=1,
            jobtype=None,
            timeField='modificationTime',
            plot='no',
            dump=None,
            tasktype=None,
            status=None,
            region=None,
            lfn=None,
            showcpu=None):
     """Display the job information: 
  <p>Sample queries:
  <ul>
   <li><a href='http://pandamon.cern.ch/jobinfo?tstart=2013-04-07 00:00&amp;tend=2013-04-07 00:03&amp;hours='>http://pandamon.cern.ch/jobinfo/?tstart=2011-06-07 00:00&amp;tend=2011-06-07 00:01</a>
    <li><a href='http://pandamon.cern.ch/jobinfo?tstart=2013-04-07 00:00&amp;tend=2013-04-07 00:03&amp;hours=&amp;processingType=simul'>http://pandamon.cern.ch/jobinfo/?tstart=2011-06-07 00:00&amp;tend=2011-06-07 00:01&[email protected]</a>
    <li><a href='http://pandamon.cern.ch/jobinfo?job=1246240909'>http://pandamon.cern.ch/jobinfo?job=1246240909</a>
   </ul>
       <hr>
       Supported query parameters are as follows:
       <table>
       <tr><td>taskid</td><td> - jobset / task id</td></tr>
       <tr><td>job</td><td> - comma separated list of jobs' PandaID</td></tr>
       <tr><td>jobtype</td><td> - select the 'jobtype' jobs: analysys, production, install, test </td></tr>
       <tr><td>tstart=None</td><td>  - start time </td></tr>
       <tr><td>tend=None</td><td>    - end time</td></tr>
       <tr><td>hours=N </td><td> - last (N hours + N days) or N hours after 'tstart'  or N hours before 'tend' </td></tr> 
       <tr><td>days=N</td><td>   - last (N hours + N days)  (see the hours parameter documentation</td></tr>
       <tr><td>lfn=</td><td>   -  Select the PanDA jobs using the files from the comma-separated lfn list </td></tr>
       <tr><td>jobparam=</td><td>The comma separated list of the <a title='Click to see the full list of the jobinfo parameters available' href='http://pandamon.cern.ch/describe?table=^JOBSARCHIVED4&doc=*'>job parameters</a>.<br> The  default should satisfy the interactive users.<br>
       One needs to customize jobparam to build his/her <a href='https://twiki.cern.ch/twiki/bin/view/PanDA/PandaPlatform#Custom_Monitoring_Client_Code'>own client</a></td></tr>
       <tr><td>limit=</d><td>   - max number of the Db records to fetch from  each job Db table. Since there are 4 tables the real number of the rows can be 4 times as many as that limit</td></tr>
       <tr><td><a title='Click to see the full list of the jobinfo parameters available' href='http://pandamon.cern.ch/describe?table=^JOBSARCHIVED4&doc=*'>PARAM</a>=value</td><td>-  comma-separated list of the value <a href ='https://twiki.cern.ch/twiki/bin/view/PanDA/PandaPlatform#API_parameter_value_format'> patterns</a> to filter the job records.<br>
       For example: <ul><li><a href='http://pandamon.cern.ch/jobinfo?processingType=simul'>http://pandamon.cern.ch/jobinfo?processingType=simul</a>
       <li> <a href='http://pandamon.cern.ch/jobinfo?limit=2&hours=4&processingType=ganga*,test'>http://pandamon.cern.ch/jobinfo?limit=2&hours=4&processingType=ganga*,test</a></ul>
       <tr><td>summary</td><td>Create the summary for the first member of the jobparam list<br>
       For example: <a href='http://pandamon.cern.ch/jobinfo?processingType=simul&jobparam=jobStatus&summary=yes'>http://pandamon.cern.ch/jobinfo?processingType=simul&jobparam=jobStatus&summary=yes</a>
       </td></tr>
       <tr><td>dump</td><td> - 'yes' - produce the flat list of job info  records suitable for CLI applications</td></tr>
       <tr><td>region</td><td> - the comma separated list of the regions </td></tr>
       <tr><td>showcpu</td><td> - 'yes' to show the CPU type used </td></tr>
       </table>
   """
     #       % { 'url' : self.server().branchUrl() }
     if job == 'undefined': job = None
     if limit == 'undefined': limit = None
     if summary == 'undefined': summary = None
     if jobparam == 'undefined': jobparam = None
     if tstart == 'undefined': tstart = None
     if tend == 'undefined': tend = None
     if tasktype == 'prod': jobtype = 'production'
     if jobtype == 'null': jobtype = None
     if jobparam == None:
         if job != None:
             jobparam = self._allparams
         else:
             jobparam = ','.join(self._colNames)
     if timeField == None or timeField == 'undefined' or timeField == '':
         timeField = 'modificationTime'
     if self.isValid(jobparam):
         self.publish(
             self.doJobs(job,
                         tstart=tstart,
                         tend=tend,
                         hours=hours,
                         days=days,
                         jobparam=jobparam,
                         limit=limit,
                         field=timeField,
                         jobtype=jobtype,
                         summary=summary,
                         minjob=minjob,
                         dump=dump,
                         region=region,
                         lfn=lfn,
                         showcpu=showcpu))
         self.publish("%s/%s" %
                      (self.server().fileScriptURL(), "jobInfoRender.js"),
                      role=pmRoles.script())
         self.publish({
             's-maxage': 600,
             'max-age': 600
         },
                      role=pmRoles.cache())
     else:
         title = 'Job data '
         if utils.isFilled(job):
             hours = None
             days = None
             title += " %s " % job
         if days != None:
             title += ' for %.2f day' % float(days)
             if days > 1: title += "s"
         if hours != None:
             if days == None: title += " for "
             title += ' %d hour' % int(hours)
             if hours > 1: title += "s"
         self.publishTitle(title)