Re: [Condor-users] Problem using schedd web service

[Peter Ledbrook <peter.ledbrook@xxxxxxxxx>]

Matthew Farrellee wrote:

> All the files that are going to be input or output (including Out/ 
> Err) should be declared.
>
> If after you declare Out and Err you still have trouble you should  
> try to add StageInStart and StageInFinish, both set to some non-zero  
> integer, to the job ad. CreateJobTemplate will add those attributes  
> for you in future versions of Condor.
>
>
> matt
>  
I'm sorry to bother again, but I'm still having a problem, although at 
least it's a different problem now! I am declaring the stdout and stderr 
files like this:


   // Declare the stdout and stderr files.
   Status retval = stub.declareFile(
       txn,
       clusterId,
       jobId,
       OUTPUT_FILENAME,
       Integer.MAX_VALUE,      // Also tried with -1
       HashType.NOHASH,
       null);
   retval = stub.declareFile(
       txn,
       clusterId,
       jobId,
       ERROR_FILENAME,
       Integer.MAX_VALUE,
       HashType.NOHASH,
       null);

and I have also included StageInStart ("10") and StageInFinish("20") in 
the JobAd. However, I am getting the following in the ShadowLog:

10/28 09:08:00 (140.0) (8981):Requesting Primary Starter
10/28 09:08:00 (140.0) (8981):Shadow: Request to run a job was ACCEPTED
10/28 09:08:00 (140.0) (8981):Shadow: RSC_SOCK connected, fd = 17
10/28 09:08:00 (140.0) (8981):Shadow: CLIENT_LOG connected, fd = 18
10/28 09:08:00 (140.0) (8981):My_Filesystem_Domain = "ixico.net"
10/28 09:08:00 (140.0) (8981):My_UID_Domain = "ixico.net"
10/28 09:08:00 (140.0) (8981):  Entering pseudo_get_file_stream
10/28 09:08:00 (140.0) (8981):  file = 
"/opt/condor-6.6.10/examples/env.remote"
10/28 09:08:00 (140.0) (8981):   Weird 0xc0a8010b
10/28 09:08:00 (140.0) (8981):   Weird 0xc0a8010b
10/28 09:08:00 (140.0) (8981):Reaped child status - pid 8983 exited with 
status 0
10/28 09:08:00 (140.0) (8981):Read: condor_restart:
10/28 09:08:00 (140.0) (8981):Read: Checkpoint file name is 
"/home/condor/spool/cluster140.proc0.subproc0"
10/28 09:08:00 (140.0) (8981):  Entering pseudo_get_file_stream
10/28 09:08:00 (140.0) (8981):  file = 
"/home/condor/spool/cluster140.proc0.subproc0"
10/28 09:08:00 (140.0) (8981):   Weird 0xc0a8010b
10/28 09:08:00 (140.0) (8981):   Weird 0xc0a8010b
10/28 09:08:00 (140.0) (8981):Read: Opened 
"/home/condor/spool/cluster140.proc0.subproc0" via file stream
10/28 09:08:00 (140.0) (8987):Failed to transfer 96 bytes (only sent -1)
10/28 09:08:00 (140.0) (8981):Reaped child status - pid 8987 exited with 
status 1
10/28 09:08:00 (140.0) (8981):Shadow: Job 140.0 exited, termsig = 9, 
coredump = 0, retcode = 0
10/28 09:08:00 (140.0) (8981):Shadow: Job was kicked off without a 
checkpoint
10/28 09:08:00 (140.0) (8981):Shadow: DoCleanup: unlinking TmpCkpt 
'/home/condor/spool/cluster140.proc0.subproc0.tmp'
10/28 09:08:00 (140.0) (8981):Trying to unlink 
/home/condor/spool/cluster140.proc0.subproc0.tmp
10/28 09:08:00 (140.0) (8981):user_time = 1 ticks
10/28 09:08:00 (140.0) (8981):sys_time = 2 ticks
10/28 09:08:00 (140.0) (8981):********** Shadow Exiting(107) **********


It looks like Condor is trying write to a directory 
(/home/condor/spool/cluster140.proc0.subproc0) as if it were a file, but 
I have no idea why. Any suggestions?

Thanks,

Peter