Re: [Condor-users] [ExternalEmail] Re: Windows jobs not running after upgrade 7.2.4 to 7.4.3

Mailing List Archives Authenticated access	UW Madison Computer Sciences Department Computer Systems Lab

condor_status -constraint "( ( ( 1024 * target.Memory ) >= 1250 ) && ( ( 1024 * ceiling(ifThenElse(JobVMMemory =!= undefined,JobVMMemory,1.220703125000000E+000)) ) >= 1250 ) )"

produces many matches. So why isn't it working?

Collector/Negotiator is linux running 7.2.3, submit node windows running 7.4.3, execute nodes

windows running 7.2.4

Maybe some relevant info that might be useful.

Following are the different requirements generated by condor for the 2 versions

RequestMemory = ceiling(ImageSize / 1024.000000)
Requirements = (Arch == "INTEL") && (OpSys == "WINNT51") && (Disk >= DiskUsage)
&& ((Memory * 1024) >= ImageSize) && (HasFileTransfer)

RequestMemory = ceiling(ifThenElse(JobVMMemory =!= UNDEFINED, JobVMMemory, Image
Size / 1024.000000))
Requirements = (Arch == "INTEL") && (OpSys == "WINNT51") && (Disk >= DiskUsage)
&& (((Memory * 1024) >= ImageSize) && ((RequestMemory * 1024) >= ImageSize)) &&
(HasFileTransfer)

We have a user who is submitting the same jobs with the same requirements _expression_

that all worked with 7.2.4 but apparently are not now with 7.4.3

requirements = (POOL == "VIC") && (Kflops > 1200000)

The jobs sit idle and never execute because no machines match, as shown by condor_q -better-analyze

013.006: Run analysis summary. Of 1229 machines,
    371 are rejected by your job's requirements
      0 reject your job because of their own requirements
      0 match but are serving users with a better priority in the pool
    839 match but reject the job for unknown reasons
     19 match but will not currently preempt their existing job
      0 match but are currently offline
      0 are available to run your job

The Requirements _expression_ for your job is:

( ( target.POOL == "VIC" ) && ( target.Kflops > 1200000 ) &&
( target.OpSys == "WINNT51" ) ) && ( target.Arch == "INTEL" ) &&
( target.Disk >= DiskUsage ) && ( ( ( target.Memory * 1024 ) >= ImageSize ) &&
( ( RequestMemory * 1024 ) >= ImageSize ) ) && ( target.HasFileTransfer )

    Condition                         Machines Matched    Suggestion
    ---------                         ----------------    ----------
1   ( ( ( 1024 * target.Memory ) >= 1250 ) && ( ( 1024 * ceiling(ifThenElse(JobVMMemory isnt undefined,JobVMMemory,1.220703125000000E+000)) ) >= 1250 ) )
                                      0                   REMOVE
2   ( target.Kflops > 1200000 )       988
3   ( target.POOL == "VIC" )          1071
4   ( target.OpSys == "WINNT51" )     1214
5   ( target.HasFileTransfer )        1223
6   ( target.Arch == "INTEL" )        1228
7   ( target.Disk >= 1500 )           1229

I've not taken a lot notice of the "extra" requirements that condor adds itself before but am

wondering about the "RequestMemory" requirement as googling seems to show that a bug

was fixed in 7.2 that made it 1024 times too big due to a mix up between Mb and Kb.

Could this still be a problem?

I've also added below the results from a condor_q -l command if that's at all relevant.

I'll keep looking into it but thought I'd try the users-list as well.

Err = "cvferr_6.txt"
LastJobStatus = 0
Out = "cvfout_6.txt"
ProcId = 6
Shortjob = TRUE
UserLog = "C:\\Users\\pok008\\TUCA\\PredSel\\jfm\\cvflog_6.txt"
JobStatus = 1
GlobalJobId = "wan110a-hr.nexus.csiro.au#13.6#1283211977"
Args = "CRNTUL1_RF.txt climate.txt pred.txt TUCjfm006 6 0 1"
ServerTime = 1283219235
ClusterId = 13
CompletionDate = 0
NTDomain = "NEXUS"
WindowsMajorVersion = 6
WindowsMinorVersion = 0
WindowsBuildNumber = 6002
WindowsServicePackMajorVersion = 2
WindowsServicePackMinorVersion = 0
WindowsProductType = 3
LocalUserCpu = 0.000000
LocalSysCpu = 0.000000
RemoteSysCpu = 0.000000
ExitStatus = 0
NumCkpts_RAW = 0
NumCkpts = 0
NumJobStarts = 0
NumRestarts = 0
NumSystemHolds = 0
CommittedTime = 0
TotalSuspensions = 0
LastSuspensionTime = 0
CumulativeSuspensionTime = 0
ExitBySignal = FALSE
CondorVersion = "$CondorVersion: 7.4.3 Aug 4 2010 BuildID: 261829 $"
CondorPlatform = "$CondorPlatform: INTEL-WINNT50 $"
Iwd = "C:\\Users\\pok008\\TUCA\\PredSel\\jfm"
MinHosts = 1
MaxHosts = 1
CurrentHosts = 0
WantRemoteSyscalls = FALSE
WantCheckpoint = FALSE
RequestCpus = 1
EnteredCurrentStatus = 1283211977
User = "pok008@xxxxxxxx"
NiceUser = FALSE
Environment = ""
JobNotification = 3
WantRemoteIO = TRUE
CoreSize = 0
Rank = ConsoleIdle
In = "/dev/null"
TransferIn = FALSE
StreamOut = FALSE
StreamErr = FALSE
BufferSize = 524288
BufferBlockSize = 32768
ShouldTransferFiles = "YES"
WhenToTransferOutput = "ON_EXIT_OR_EVICT"
TransferFiles = "ALWAYS"
TransferInput = "CRNTUL1_RF.txt,climate.txt,pred.txt"
ImageSize_RAW = 1204
ExecutableSize_RAW = 1204
ExecutableSize = 1250
DiskUsage_RAW = 1374
DiskUsage = 1500
RequestMemory = ceiling(ifThenElse(JobVMMemory =!= UNDEFINED, JobVMMemory, Image
Size / 1024.000000))
RequestDisk = DiskUsage
Requirements = ((POOL == "VIC") && (Kflops > 1200000) && (OpSys == "WINNT51")) &
& (Arch == "INTEL") && (Disk >= DiskUsage) && (((Memory * 1024) >= ImageSize) &&
((RequestMemory * 1024) >= ImageSize)) && (HasFileTransfer)
JobLeaseDuration = 300
PeriodicHold = FALSE
PeriodicRelease = FALSE
PeriodicRemove = FALSE
> == 0)
LeaveJobInQueue = FALSE
Owner = "pok008"
JobPrio = 3000
ImageSize = 1250
QDate = 1283211977
RemoteUserCpu = 0
RemoteWallClockTime = 0
Cmd = "C:\\Users\\pok008\\cvcluster_select.exe"
JobUniverse = 5

Mailing List Archives

Authenticated access

Re: [Condor-users] [ExternalEmail] Re: Windows jobs not running after upgrade 7.2.4 to 7.4.3