CheckIdl

From PTAGISWiki

Jump to: navigation, search

Useful aliases as user ptagdev

Click the discusion tab to see what's going on...

Add discussion to the top of the discussion page.


jld - Show start and finish times of most recent p3_daily jobs
see_idl - shows IDL files to load
fc_idl - shows number of files to load
rc_idl - shows number of records to load

You may want to use the work-script I've been compiling today as I ride herd on IDL. It's on 'sockeye' at /home/ptagdev/ptagis3/sys_config_chores/ProbLog. The file is:

   ProblemDescription_ProcessCreationSpike.txt

In the file you'll find step-by-step the commands I've issued so far today. Also in that directory is the file I used when catching up from the scheduling problem that occurred Monday night. It's called

   ProblemDescription_LongRunningP3_daily_4.txt

It may be useful to have a look at this file, to get an overview of what I did on Tuesday - which will be practically identical to things you or someone will need to do this afternoon.

Sorry to leave when things are balled up like this. I'll give you a call before I take off, so you can ask questions that may come up when you look at these work-script files.


                       Pacific States Marine Fisheries Commission
                           PTAGIS3 Interrogation Data Loader

                             Long-Running 'p3_daily_4' Job

                                 Ver 1.0 09 May 2006
                                (Last Update:  11:30)

History:
---------    --------------    -------------------------------------------------
09 May 06    Doug Clough       Creation, at about 07:00 PDT


==OVERVIEW==

'p3_daily_2.sh' stopped because p3_daily_4.sh launched at 18:15 PST ran for 10-
minutes beyond the p3_daily_2.sh launch time of 00:15 PST. This meant 90-plus
files (over 235,000 rows) would have to be processed by the p3_daily_3.sh run
launched at 06:15 PST on 09 May.

As detailed below, the long run-time had several causes: First, this is a very busy
time of year. On top of that, large quantities of PIT-tagged fish are being released
at GOJ and GRJ this year, resulting in unusually large numbers of interrogations at
these sites. Further, two acclimation-pond interrogation files (from CCP and GRP)
were placed in the input queue yesterday, just prior to the 18:15 PST IDL launch.
Similarly, TWX interrogation files spanning 2 thru 9 May had been presented for
loading.  Finally, yesterday being a Monday, CEA was tasked with producing 'weekly'
and '2-week' aggregations during the p3_daily_4.sh run in question.

Without the CCP and GRP files, IDL would have finished more than 20-minutes earlier;
without the TWX files, it would have finished about 15-minutes earlier: In either
case, the scheduling problem would not have arisen.


==INTERVENTION==

  Done  1. Disable 'cron' processing in the 'pittag' account ...
           > cd /home/pittag/bin
           > su - pittag
     OK    > crontab -r

           DPC  Tue May  9 08:16:25 PST 2006

  Done  2. Manually adjust 'cronfile' to launch another p3_daily_3.sh just as
           soon as the one launched at 06:15 PST completes
     OK    > cp -p cronfile cronfile.09may06

           DPC  Tue May  9 11:01:02 PST 2006

           As 'ptagdev', watch for p3_daily_3 completion expected ~12:30 PDT ...
           > jld

           As 'pittag' ...
           > cd /home/pittag/bin
     OK    > vi cronfile.09may06
             Re-schedule to launch p3_daily_3.sh at 12:45 PDT

     OK    > crontab cronfile.09may06

           Once the job has launched ...
     OK    > crontab -r

           DPC  Tue May  9 11:45:51 PST 2006

           As 'ptagdev', resume watching for p3_daily_3 completion ...
           > jld

           As 'pittag' ...
           > cd /home/pittag/bin
     OK    > vi cronfile.09may06
             Re-schedule to launch p3_daily_3.sh at 15:00 PDT

     OK    > crontab cronfile.09may06

           Once the job has launched ...
     OK    > crontab -r

           DPC  Tue May  9 14:07:55 PST 2006

           As 'ptagdev', resume watching for p3_daily_3 completion ...
           > jld


  Done  3. Re-enable 'cron' processing on the normal schedule ...
           > cd /home/pittag/bin
           > su - pittag
     OK    > crontab -r
     OK    > crontab cronfile
     OK    > crontab -l

           DPC  Tue May  9 15:34:15 PST 2006

-v-v-v-v-v-v-v-v-v-v-v-v Extra info in case it's needed -v-v-v-v-v-v-v-v-v-v-v-

           'snapckp' is launched from 'root' crontab at 23:00

           Here's how to enable / disable snapckp ...
           > cat /home/pittag/etc/SNAPCKP_FLAG
             # The VALUE set below controls invocation of ^snapckp^  from
             # the root crontab. VALUE may be either RUN or SKIP.
             VALUE: RUN

-^-^-^-^-^-^-^-^-^-^-^-^ Extra info in case it's needed -^-^-^-^-^-^-^-^-^-^-^-
   IDL

      06:38 PST    91 files   235802 rows
                                           10,471 rows / 25 min = 418 rows / min
      07:03 PST    76 files   225331 rows
                                           15,863 rows / 22 min = 721 rows / min
      07:25 PST    66 files   209468 rows
                                           77,073 rows / 93 min = 829 rows / min
      08:58 PST    44 files   132395 rows
        At 800 rows / min, projected completion time is 2hr 45 min, or
        11:45 PST --> 12:45 PDT


ANALYSIS

IDL

   Dave mentioned that he submitted two acclimation pond interrogation files
   at the end of the day yesterday; maybe these had an impact. I checked it
   out:
      CCP  14,987 rows, 23-minutes elapsed time
      GRP     439        1-minute
   Yes, without these, p3_daily_4.sh would have finished just before midnight;
   15-minutes before the launch of p3_daily_2.sh

CEA

CEA_reportYTDData.pl Begins at Tue May  9 00:24:04 PST 2006
CEA_reportYTDData.pl Ends   at Tue May  9 00:25:15 PST 2006

CEA_reportHistoricalData.pl Begins at Tue May  9 00:22:56 PST 2006
CEA_reportHistoricalData.pl Ends   at Tue May  9 00:23:58 PST 2006


CEA_processHistoricalData.pl Begins at Mon May  8 19:59:32 PST 2006
   daily

      Loading data   at Mon May  8 20:19:42 PST 2006

CEA_processHistoricalData.pl Begins at Mon May  8 20:20:27 PST 2006
   weekly

CEA_processHistoricalData.pl Begins at Mon May  8 22:01:19 PST 2006
   2-week
     GOJ
      Running report at Mon May  8 22:03:28 PST 2006
      Loading data   at Mon May  8 22:54:34 PST 2006
     GRJ
      Running report at Mon May  8 22:54:42 PST 2006
      Loading data   at Mon May  8 23:10:41 PST 2006
      Running report at Mon May  8 22:54:42 PST 2006
      Loading data   at Mon May  8 23:10:41 PST 2006
     ICH
      Running report at Mon May  8 23:10:47 PST 2006
      Loading data   at Mon May  8 23:15:11 PST 2006
     JDJ
      Running report at Mon May  8 23:15:24 PST 2006
      Loading data   at Mon May  8 23:21:40 PST 2006
     LMJ
      Running report at Mon May  8 23:21:43 PST 2006
      Loading data   at Mon May  8 23:50:29 PST 2006
     MCJ
      Running report at Mon May  8 23:52:28 PST 2006
      Loading data   at Tue May  9 00:21:45 PST 2006
     PRO
      Running report at Tue May  9 00:21:58 PST 2006
      Loading data   at Tue May  9 00:22:18 PST 2006
     WEA
      Running report at Tue May  9 00:22:51 PST 2006
      Loading data   at Tue May  9 00:22:53 PST 2006


|--------|---------|---------|---------|---------|---------|---------|---------|
Personal tools