Calendar buckets out by 90% (1296 minutes)

[Chris]

Hi,

I am learning frePPLe (0.9.2 and tries 0.9.3) and am having some trouble.

I am in timezone GMT+8 and have set my TIME_ZONE = 'Asia/Kuala_Lumpur' in the settings.py

 

If i use the initial_data and the Working Days calendar, it works ok.  but as soon as I try to create a new calendar my work days dont match.

For example, If i manually create a calendar and buckets for a work week (say Mon-Sat, 09:00-16:59:00) Under recource plan, will show the following values:

 

Friday: Available =1, Unavailable = 0

Saturday: Available = 0.1, Unavailable = 0.9

Sunday: Available - 0.9, Unavailable = 0.1

 

Its like Sunday only gets 10% (144 minutes)

 

Even after this if i switch back to the original calendar, it still gets screwed up and wont get back to sunday = 100% Unavailable until i reset the DB and start again.

 

I have checked the calendar and calendarbucket, as well as the common_calendar and common_calendarbucket tables and the data seems fine.

 

Any ideas?

 

Thanks,

Chris

[Johan De Taeye]

Chris,

 

I guess the timezones are indeed the cause of the problem.  
Daylight Saving Time creates a 23-hour day or a 25-day twice a year, but this can account only for a 1 hour difference.

 

To analyze the problem:

-        Which database are you using?

-        Imagine you have a demand with a due datetime of March 29 at 17h.
Do you see the same values in the UI as directly in the database?

-        If you run frepple in debug mode (by editing the file execute/commands.py – setting loglevel to 2), do you find the very same due date&time in the log file?

If all these tests are correct, then dates & times are passed correctly between the planning engine, the database and the user interface.  And the problem might be a bug in the planning engine.

 

Johan

--
You received this message because you are subscribed to the Google Groups "frePPLe users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to frepple-user...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Chris]

Hi Johan,

thanks for replying.

the resource detail seems correct:

1	TW17970210	1,600.00	2013-04-15 09:06:40	2013-04-15 14:00:00	False	0.00	0
2	TW17970210	2,600.00	2013-04-13 09:03:20	2013-04-13 17:00:00	False	0.00	0
3	TW17970210	2,600.00	2013-04-12 09:03:20	2013-04-12 17:00:00	False	0.00	0
4	TW17970210	2,600.00	2013-04-11 09:03:20	2013-04-11 17:00:00	False	0.00	0
5	TW17970210	600.00	2013-04-10 15:10:00	2013-04-10 17:00:00	False	0.00	0

Sunday is the 14th, and should not be worked.

But in the Resource plan view the values are:

4/9/2013	4/10/2013	4/11/2013	4/12/2013	4/13/2013	4/14/2013	4/15/2013
0.3	0.3	0.3	0.3	0.1	0.3	0.3
0	0	0	0	0	0	0
0	0	0	0	0	0	0
0	0.3	0.3	0.3	0.1	0.2	0
0%	99.31%	99.31%	99.31%	100%	79.28%	0%

I will look at the log file, but the data in the DB seems ok.

Thanks,

Chris

[Johan De Taeye]

Chris,

 

Can you show the records you find in the out_resourceplan table in the database?

They are expected to be a summary in daily buckets of the resource detail, and is the input for the resource plan view.

[Tim Sharpe]

Johan,

 

We are noticing something similar to Chris' issue.

 

Five day work week calendar (Mon through Fri) makes Saturday unavailable, but does partial planning on Friday and Sunday:

 

Example from Resource report:

 

Friday 2013-03-01

  Available  10.1 hours / Unavailable 13.9 hours

Saturday 2013-03-02
  Available  0 hours / Unavailable 24 hours

 

Sunday 2013-03-03

  Available  9.9 hours / Unavailable 14.1 hours

 

We are only interested in capacity planning, and since the total number of hours seems to be okay, I wasn't too concerned about this.

 

Tim

[Johan De Taeye]

 

Chris, Tim,

 

I tried reproducing the problem, but wasn’t able to.

Also tried changing the timezone on my machine but that isn’t the problem apparantly.

 

Only possibility I see is that the resource and the operations are in different locations, with different availability calendars.

If that isn’t the case in your model, I’ld need more info to analyze the issue.

 

Regards,

[jdetaeye]

This commit should resolve the issue...
http://sourceforge.net/p/frepple/code/1865/

Johan

> execute/commands.py - setting loglevel to 2), do you find the very same due

> date&time in the log file?
>
> If all these tests are correct, then dates & times are passed correctly
> between the planning engine, the database and the user interface.  And the
> problem might be a bug in the planning engine.
>
> Johan
>

> email to frepple-user...@googlegroups.com <javascript:> .
> For more options, visithttps://groups.google.com/groups/opt_out.

[Chris]

Thanks Johan,

I will deploy and test with my users.

Cheers,

Chris