Re: [HTCondor-users] Jobs Stuck Due to RequestDisk = undefined

["ëêë" <geonmo@xxxxxxxxxxx>]

Hello, Eraldo.

I think we already have an answer to your question.

```[0]           0  TARGET.Arch == "X86_64"```

I see that you don't have any servers with x86_64 architecture, so I'm wondering if your WNs are using ARM CPUs.

You might need to check that.

Regards,

-- Geonmo

ââââââ ìë ëì ââââââ

ëëìë : Eraldo Jr <eusoueraldo@xxxxxxxxx>

ëëìë : HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>

ëìëì : 2025-04-15 (í) 10:59:21

ìë : [HTCondor-users] Jobs Stuck Due to RequestDisk = undefined

Dears,

First of all:

$CondorVersion: 24.4.0 2025-02-02 BuildID: 784192 PackageID: 24.4.0-1 GitSHA: 6f17b75e $
$CondorPlatform: x86_64_AlmaLinux9 $

I am encountering an issue where several jobs are failing to match available slots, despite sufficient disk space being reported. The problem seems to be related to how RequestDisk is being evaluated in conjunction with dynamically allocated slots.

Issue Summary:

• Jobs are specifying RequestDisk = DiskUsage, but some still appear to have RequestDisk as undefined or fail to match slots correctly.
• The available dynamic slots report ample disk space (TotalDisk ≈ 800GB per slot).
• However, certain jobs have high DiskUsage values (e.g., DiskUsage = 11GB), and they are failing to find a suitable match.
• Running condor_ce_q -better-analyze shows that the disk requirements are preventing allocation.

=======================================

The Requirements _expression_ for job 9867.000 is

    (TARGET.Arch == "X86_64") && (TARGET.OpSys == "LINUX") && (TARGET.Disk >= RequestDisk) &&

    (TARGET.Memory >= RequestMemory) && (TARGET.HasFileTransfer)

    [0]    : TARGET.Arch == "X86_64"

    [1]    : TARGET.OpSys == "LINUX"

    [2]    : [0] && [1]

    [3]    : TARGET.Disk >= RequestDisk

    [4]    : [2] && [3]

    [5]    : TARGET.Memory >= RequestMemory

    [6]    : [4] && [5]

    [7]    : TARGET.HasFileTransfer

    [8]    : [6] && [7]

Job 9867.000 defines the following attributes:

    DiskUsage = 40

    ImageSize = 40

    RequestDisk = undefined (kb)

    RequestMemory = ifthenelse(MemoryUsage =!= undefined,MemoryUsage,(ImageSize + 1023) / 1024) (mb)

The Requirements _expression_ for job 9867.000 reduces to these conditions:

        Slots

Step   Matched  Condition

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

[0]           0  TARGET.Arch == "X86_64"

[1]           0  TARGET.OpSys == "LINUX"

[3]           0  TARGET.Disk >= RequestDisk

[5]           0  TARGET.Memory >= RequestMemory

[7]           0  TARGET.HasFileTransfer

=======================================

[root@ce04 ~]# condor_ce_q -l 9867 |grep Disk|more

DiskUsage = 40

DiskUsage_RAW = 39

RequestDisk = DiskUsage

Requirements = (TARGET.Arch == "X86_64") && (TARGET.OpSys == "LINUX") && (TARGET.Disk >= RequestDisk) &

& (TARGET.Memory >= RequestMemory) && (TARGET.HasFileTransfer)

=======================================

[root@ce04 ~]# _val -dump | grep REQUEST

e_config_val -dump | grep DiskUsage-bash: _val: command not found

[root@ce04 ~]# condor_ce_config_val -dump | grep DiskUsage

JOB_DEFAULT_REQUESTDISK = DiskUsage

SCHEDD_ROUND_ATTR_DiskUsage = 25%

SYSTEM_STARTD_JOB_ATTRS = ImageSize, ExecutableSize, JobUniverse, NiceUser, CPUsUsage, ResidentSetSize, ProportionalSetSizeKb, MemoryUsage, DiskUsage, ScratchDirFileCount

=======================================

[root@ce04 ~]#  condor_status -long | grep DiskUsage|more

DiskUsage = 236135

DiskUsage = 275046

DiskUsage = 249114

DiskUsage = 480483

DiskUsage = 377061

DiskUsage = 272785

DiskUsage = 306690

DiskUsage = 321216

DiskUsage = 294344

DiskUsage = 156775

DiskUsage = 256135

DiskUsage = 300316

DiskUsage = 306545

DiskUsage = 221228

DiskUsage = 311412

=======================================

I appreciate any guidance you can provide!

Best regards,
Eraldo