nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

getting parent jobs of stalled tasks

CRB

Hello,
i run ts_diagnostics_win on TS 5.5.2 on w2k. and it showed this:
--------------------------------------------------------------------
Stalled tasks: You have 653 tasks that have stalled. These are submit or update tasks that have stalled because of conflicts or are cgi tasks that are ready to be run. If there are many of either you may want to look at and remove all jobs associated with these tasks. You should call technical support for information on these if there are many to find out how to prevent them in the future. You can also use the list of these this program generated, /tmp/stalledtasks.txt, and run something like this on them: ----- # getParentjob.ipl use strict; use TeamSite::WFtask(); ### get the task id and return the parent..
my $strFile = $ARGV[0];
open( F, ">$strFile" ) || die "can't find file "; my @arrTasks = ; close ( F ); foreach ( @arrTasks ){ chomp( $_ ); } foreach my $iTaskid ( @arrTasks ){ my $oTask = new TeamSite::WFtask( $iTaskid ); print $oTask->GetWorkflowId() . "\n"; } 1;
------------------------ Then run: iwperl getParentjob.ipl stalled.txt Then run iwrmjob on each of the jobs returned
-------------------------------------------------------

so to remove the stalled tasks i did this:
I created this script and named it getParentjob.ipl:
----------------------------------------
# getParentjob.ipl
use strict;
use TeamSite::WFtask();

### get the task id and return the parent..
my $strFile = $ARGV[0];
open( F, ">$strFile" ) || die "can't find file "; my @arrTasks = ; close ( F );
foreach ( @arrTasks ){ chomp( $_ ); }
foreach my $iTaskid ( @arrTasks )
{
my $oTask = new TeamSite::WFtask( $iTaskid );
print $oTask->GetWorkflowId() . "\n";
}
1;
------------------------------------------------

Then I will run it with this command:
iwperl getParentjob.ipl stalledtaks.txt

Then I will run iwrmjob on the output,

Is this correct?

when i run "iwperl getParentjob.ipl stalledtaks.txt" it just clears the stalledtasks.txt file and it doesnt print out anything.. i am debuggin wht i am doing right now, but i need help confirming that i am doing is correct?

Find more posts tagged with

Comments

CRB

i found that my issue here, that the open( F, ">$strFile" ) || die "can't find file "; doesnt put anything in my @arrTasks = ; it just clears the .txt file... i have to look at my perl books Smiley Very Happy

any help here?

CRB

it seems like it takes along time to getworkflowid for each stalled taskid... but so far i noticed that every workflowid = stalledtaskid-1 like 1624 = 1625 -1 ... i am not sure if i wanna wait for this till it finishs... but i might... and eventually i will run iwrmjob on every id i get in this list... but i am affraid these ids will be of some jobs that are needed right now by the dev team... how can i make sure that the job i am deleting is old? i see the activation time is in seconds in those jobs for example "1079479003"... if i can confirm that that date is an old date i can for sure go ahead and delete it...

Adam Stoller

it seems like it takes along time to getworkflowid for each stalled taskid... but so far i noticed that every workflowid = stalledtaskid-1 like 1624 = 1625 -1 ... i am not sure if i wanna wait for this till it finishs... but i might... and eventually i will run iwrmjob on every id i get in this list... but i am affraid these ids will be of some jobs that are needed right now by the dev team... how can i make sure that the job i am deleting is old? i see the activation time is in seconds in those jobs for example "1079479003"... if i can confirm that that date is an old date i can for sure go ahead and delete it...

Take the activation time and use this command-line script with it:

iwperl -le "print scalar localtime(1079479003);"

CRB

thanks ghoti.
i had to modify the code little to get it working... i need the workflowids to be printed to a list that i can use later with iwrmjob but there is one part that i couldnt get working, which is getting the: $oTask->GetWorkflowId(); to print out to a file...
this is my modified code:
---------------------------------------
# getParentjob.ipl
use strict;
use TeamSite::WFtask();

### get the task id and return the parent..
my $strFile = "stalledtasks.txt"; #$ARGV[0];
open( F, "$strFile" ) || die "can't find file ";
my @arrTasks = ;
close ( F );

foreach ( @arrTasks ){ chomp( $_ ); print "chomped $_ \n";}

print "\n****\n inside 2nd loop \n****\n";

#my $strFile1 = "stalled.txt";
#open F1, "> stalled.txt" or die "can't find file "; #<--the file gets created everytime

foreach my $iTaskid ( @arrTasks )
{
my $oTask = new TeamSite::WFtask( $iTaskid );
my $id = $oTask->GetWorkflowId();
print $id ."\n"; #<-- this prints out fine in the command print window
print F1 $id ."\n"; #<-- this doesnt print anything the created file says 0k size...
}

close ( F1 );
print " finished\n";
1;
------------------------
i tried the iwperl getparentjob.ipl >output.txt it prints out starting from the first loop "chomped" but it doesnt print all the records.. the output.txt always stayed in 5k size and always stoped at the same record " i have 653 record".

i got somebody else to look at the code, and he confirmd that my syntax is correcty, and it couldnt tell why it is not printin $oTask->GetWorkflowId();to a file

page.png

Adam Stoller

thanks ghoti.
i had to modify the code little to get it working... i need the workflowids to be printed to a list that i can use later with iwrmjob but there is one part that i couldnt get working, which is getting the: $oTask->GetWorkflowId(); to print out to a file...
this is my modified code:
---------------------------------------
# getParentjob.ipl
use strict;
use TeamSite::WFtask();

### get the task id and return the parent..
my $strFile = "stalledtasks.txt"; #$ARGV[0];
open( F, "$strFile" ) || die "can't find file ";
my @arrTasks = <F>;
close ( F );

# The above opens "stalledtasks.txt" in whatever the current working directory is (BAD - should specify a path here) and reads the contents into an array with each element of the array corresponding to each line in the file.
# How was the file created? Are you sure the format is correct? Presumably it is a list of taskids, one per line.

foreach ( @arrTasks ){ chomp( $_ ); print "chomped $_ \n";}

print "\n****\n inside 2nd loop \n****\n";

#my $strFile1 = "stalled.txt";
#open F1, "> stalled.txt" or die "can't find file "; #<--the file gets created everytime

# The above was commented out in your post - meaning that you never opened the file stalled.txt (BAD - no path provided) for writing - and thus your later attempts to print into, and close, the file handle F1 are probably generating errors somewhere.

foreach my $iTaskid ( @arrTasks )
{
my $oTask = new TeamSite::WFtask( $iTaskid );
my $id = $oTask->GetWorkflowId();
print $id ."\n"; #<-- this prints out fine in the command print window
print F1 $id ."\n"; #<-- this doesnt print anything the created file says 0k size...
}

close ( F1 );
print " finished\n";
1;
------------------------
i tried the iwperl getparentjob.ipl >output.txt it prints out starting from the first loop "chomped" but it doesnt print all the records.. the output.txt always stayed in 5k size and always stoped at the same record " i have 653 record".

Not sure what the issue is there - but I suggest fixing the above issues first before worrying about it.

i got somebody else to look at the code, and he confirmd that my syntax is correcty, and it couldnt tell why it is not printin $oTask->GetWorkflowId();to a file

CRB

# The above opens "stalledtasks.txt" in whatever the current working directory is (BAD - should specify a path here) and reads the contents into an array with each element of the array corresponding to each line in the file.

this part was working fine... i just made sure i have the file in the same folder where the script is... it might be a windows thing, but it did read the file and got the taskids from it.

[color=blue]#
# How was the file created? Are you sure the format is correct? Presumably it is a list of taskids, one per line.# The above was commented out in your post - meaning that you never opened the file stalled.txt (BAD - no path provided) for writing - and thus your later attempts to print into, and close, the file handle F1 are probably generating errors somewhere.Not sure what the issue is there - but I suggest fixing the above issues first before worrying about it.

the# was typo, i didnt have that in my script, so the file was created since it wasnt commented out.

i got this worked out... and i finished deleting all the stalledtasks that the ts_diagnostic_win tool detected.

i am wondering if that will make significant change to the performance.. i had 653 stalled tasks.

also it seems like our production team never closes tasks.. when i go to TS and click "see all jobs", it will take few mins and then i get 350 pages of workflows, each page has about 10 workflow all under "assigned status" since 2004... i wonder if that affects our performance? should we delete those workflows?

Adam Stoller

...i got this worked out... and i finished deleting all the stalledtasks that the ts_diagnostic_win tool detected.

i am wondering if that will make significant change to the performance.. i had 653 stalled tasks.

also it seems like our production team never closes tasks.. when i go to TS and click "see all jobs", it will take few mins and then i get 350 pages of workflows, each page has about 10 workflow all under "assigned status" since 2004... i wonder if that affects our performance? should we delete those workflows?

350 pages where the default is I think something like 25 entries per page ... that comes out to 8750 jobs - yeah, that might hurt performance (There used to be a performance degradation somewhere around 2000 - or maybe 20000 - tasks, not sure if that's still an issue....)

I'd try to get them cleaned up, find out why they never end those jobs (if they need reporting then they should modify the workflows to dump reporting data prior to ending - not keep the jobs lying around forever)

CRB

i went to the job workflow that they never delete and i got this:
the last step in the workflow is:

-------------------------
endtask name="end" description="The End.">
__INSERT__($activation_end_task);
/endtask>
---------------

and $activation_end_task has this:
--------------------------------

 i am having hard time because of "<", how do u skipp it on herE?
my($activation_end_task) = <<"END_OF_TEXT";
 activation>
  or>
   pred v="SubmitIt"/>
   pred v="LogIt"/>
   pred v="TransformIt"/>
  /or>
 /activation>
END_OF_TEXT

my($log_task) = <<"END_OF_TEXT";
 externaltask name="LogIt" 
             start="f" 
             owner="__TAG__('iw_user');" 
             description="Capture return codes from TransformIt task."
        retry = "t">
        areavpath v="__TAG__('iw_workarea');"/>
        successors>
            successorset description="onCompleteLog">
                succ v="end"/>
            /successorset>
        /successors>
  command v="d:\\iw-home\\iw-perl\\bin\\perl d:\\iw-home\\iw-perl\\lib\\mail\\send_transform_log.pl"/>
        activation>
            pred v="TransformIt"/>
        /activation>
    /externaltask>
END_OF_TEXT

-----------------------------

i see the last thing they do, is log the workflow... but after there i dont see anywhere where they kill the workflow? do you agree with me?

i am wondering if i should add a step on here that deletes the workflow after it is logged...

Adam Stoller

Either use [html]...[/html] tags or type <<x>tag>...<<x>/tag> within [code]...[/code] tags to avoid interpretation of the XML tags.

i went to the job workflow that they never delete and i got this:
the last step in the workflow is:

-------------------------
<endtask name="end" description="The End.">
 __INSERT__($activation_end_task);
    </endtask>
---------------

and $activation_end_task has this:
--------------------------------
 i am having hard time because of "<", how do u skipp it on herE?
my($activation_end_task) = <<"END_OF_TEXT";
 <activation>
  <or>
   <pred v="SubmitIt"/>
   <pred v="LogIt"/>
   <pred v="TransformIt"/>
  </or>
 </activation>
END_OF_TEXT

my($log_task) = <<"END_OF_TEXT";
 <externaltask name="LogIt" ...>
        <areavpath v="__TAG__('iw_workarea');"/>
        <successors>
            <successorset description="onCompleteLog">
                <succ v="end"/>
            </successorset>
        </successors>
        ...
    </externaltask>
END_OF_TEXT
-----------------------------

i see the last thing they do, is log the workflow... but after there i dont see anywhere where they kill the workflow? do you agree with me?

i am wondering if i should add a step on here that deletes the workflow after it is logged...

The externaltask "LogIt" calls the successor "end" assuming that the script associated with the externaltask performs it's callback correctly.

The endtask "end" can be activated if any of the three predecessors ("SubmitIt", "LogIt", or "TransformIt") correctly transitions to it.