Some middle ware (for example GT5) require that a separate permanent socket will be opened and maintained until the termination of the associated job instance.
A user submitting a high number of PS jobs may cause DOS (denial of service) problems as the capacity of opened sockets of a destination resource may be exhausted.
Therefore the number of the opened sockets of such DCI kind must be limited and the eventual starvation of jobs of other users must be prevented.
The introduction of three general parameters is suggested by which the gateway (more precisely the DCI-BRIDGE) administrator can trim the job flow by automatic suspension and interruption:
Two kind of errors found during job submission (together 69 error out of 1000 identical job submission):
:
1: "DCI-BRIDGE" error (29 pieces)
Error:
Reply wait timeout. (error code 4)
The job's RSL was: &(executable=./wrapper.sh)(directory=537deb24-9999-496f-8be1-49ff071155ae/)(stdout=stdout.log)(stderr=stderr.log)(jobtype=single)Error:
Server refused performing the request. Custom message: Server refused creating directory (error code 1) [Nested exception message: Custom message: Unexpected reply: 500-Command failed : globus_gridftp_server_file.c:globus_l_gfs_file_mkdir:554:
500-System error in mkdir: File exists
500-A system call failed: File exists
500 End.]
The job's RSL was: &(executable=./wrapper.sh)(directory=537deb24-9999-496f-8be1-49ff071155ae/)(stdout=stdout.log)(stderr=stderr.log)(jobtype=single)
2: "ngs.wmin.ac.uk:2119/jobmanager-pbs" error (40 pieces)
gzip: stdin: unexpected end of file
tar: Child returned status 1
tar: Error exit delayed from previous errors
chmod: cannot access `./execute.bin': No such file or directory
././wrapper.sh: line 106: ./execute.bin: No such file or directory
tar: OUTPUT*: Cannot stat: No such file or directory
tar: Error exit delayed from previous errors
Can not copy the Output file:OUTPUT