AWS Thinkbox Discussion Forums

Deadlinecommand statistics

Deprecated Deadline 6 Bug Reports
#1

Hello!

I’ve been trying to write a script using “deadlinecommand” output but came across an obstacle:

From the manual: (btw. there is typo -> “user” should be “username”)

deadlinecommand -getjobsfilter user=jsmith status=completed

However it does not return proper jobs. Using only one of these is working fine, but combining the is not giving the correct results, as if it was taking only one parameter into account. I’ve tried different combinations/syntax, without luck however.

Cheers!
Wojtek

#2

Hi Wojtek,

I’m guessing you want jobs that are complete AND match that user? If so, you’ll want the “-getjobsfilterand” option. The “-getjobsfilter” is returning jobs that are complete OR jobs that match that user.

deadlinecommand -getjobsfilter user=jsmith status=completed

Also, thanks for catching that typo. We’ll get that fixed in the updated 5.1 docs.

Cheers,

  • Ryan
#3

We’ll also update the command to show the proper command. We’ll actually include examples of both. :slight_smile:

#4

Ok, it’s working, thanks! :wink:

Cheers!
Wojtek

#5

uh, one more:

deadlinecommand -setjobsetting 003_050_000_6d832894 machinelimit 10

throws:

Error: Cannot set setting machinelimit this way (System.Exception)

#6

You want to use “SetJobMachineLimit” here.

C:\Users\Ryan>deadlinecommand help setjobmachinelimit

SetJobMachineLimit            Sets the machine limit for the job
  [Job ID(s)]                 The job ID, or a list of job IDs separated by
                              commas
  [Limit]                     The new limit
  [Listed]                    The listed slaves
  [Whilelist Flag]            true/false
  [Limit Progress]            The limit progress (optional)

Cheers,

  • Ryan
#7

Ah, there it is.

Many thanks!
Cheers!
Wojtek

#8

Ok, one more:

After submitting a job (houdini) with another depending on it as “pending” (mantra) when the first job finishes it’s tasks and move on to release the “pending one” gathering the job statistics through “getjobsfilter” the previously pending job has still the status of “pending” instead of “active”.

#9

I can’t seem to reproduce this one. How/when are you gathering these stats after the pending job is released?

#10

Hi Ryan!

We launch our custom command line Houdini job to make the IFD files for our custom command line Mantra plugin and like in Your plugins it submits hbatch job along with pending mantra one. When you “-getjobsfilter” of “active”, it returns properly the Houdini one, and “-getjobsfilter” of “pending” it returns properly Mantra one. But as soon as the Mantra is released by the completion of the Houdini, and it changes it’s status from “pending” to “active” in the Monitor, in “deadlinecommand -getjobsfilter” it is still shown as “pending”. It happens till the Mantra job is finished and changes it’s status to “complete”.

We’re querying this every 15 seconds through custom script, which is calling “deadlinecommand”. And we have Pulse running. It’s a setup for balancing the priorities of multiple users.

#11

Would it be possible to post your script? If the script just loops indefinitely, it may be a caching problem.

Thanks!

  • Ryan
#12

Umm… i’ll talk to my colleague. But it even happens just through using this command in the terminal during those jobs.

Maybe it’s a Fedora issue, or that we havent upgraded it from RC2 to final yet.

Privacy | Site terms | Cookie preferences