cpus and sge error

[ylluX]

Hi，I have the following scripts:

test.nf

#!/usr/bin/env nextflow

nextflow.enable.dsl=2


process test1 {

    cpus = 1

    input:
        val name
    output:
        val name
        path "${name}.test1.txt"
    script:
    """
    date > "${name}.test1.txt"
    sleep 10
    date >> "${name}.test1.txt"
    """

}

process test2 {

    cpus = 2
    executor 'sge'

    input:
        val name
        path test1_out
    output:
        path "${name}.test2.txt"
    script:
    """
    date > "${name}.test2.txt"
    sleep 5
    date >> "${name}.test2.txt"
    echo ${task.cpus} >> "${name}.test2.txt"
    """
}

process test3 {

    input:
        path test2_outs
    output:
        path "test3.txt"
    script:
    """
    date > test3.txt
    sleep 5
    date >> test3.txt
    """
}


workflow {
    names = channel.fromList(['A', 'B', 'C', 'D', 'E'])
    test1(names)
    test2(test1.out[0], test1.out[1])
    test3(test2.out.collect())
}

test.config

executor {
    $local {
        cpus = 3
    }
    $sge {
        queueSize = 3
        queue = 'all.q'
    }
}

process.executor = 'sge'

When I execute the command （nextflow -C test.config run test.nf）I get an error：

N E X T F L O W  ~  version 22.04.3
Launching `test.nf` [intergalactic_goodall] DSL2 - revision: 27e592bff8
executor >  sge (5)
[3e/5f20b5] process > test1 (4) [ 60%] 3 of 5
[-        ] process > test2     [  0%] 0 of 3
executor >  sge (5)
[de/974b8b] process > test1 (2) [100%] 3 of 3
[30/da87d1] process > test2 (1) [ 33%] 1 of 3, failed: 1
[-        ] process > test3     -
Error executing process > 'test2 (1)'

Caused by:
  Failed to submit process to grid scheduler for execution

Command executed:

  qsub -terse .command.run

Command exit status:
  -

Command output:
  (empty)

Work dir:
  /test/nextflow/work/30/da87d19d171d770a553751516491df

Tip: when you have fixed the problem you can continue the execution adding the option `-resume` to the run command line

When I execute the command (qsub .command.run) in the directory（/test/nextflow/work/30/da87d19d171d770a553751516491df）, I get an error:

Unable to run job: "job" denied: use parallel environments instead of requesting slots explicitly.
Exiting.

I found that the error was caused by "#$ -l slots=2" in the .command.run.

How can I delete -l slots=2 without modifying cpus=1?

[bentsherman]

Not an SGE user but I think you need to use the penv directive: https://nextflow.io/docs/latest/process.html#penv