Just some additional info, confirm that TMPDIR exists in job environment and the sge_execd does create the TMPDIR path on the designate file system. Here is my test:
$ cat sge_dd_allq.sh #!/bin/bash # # It prints the actual path of the job scratch directory. # #$ -j y #$ -cwd #$ -N dd_allq #$ -l mem_requested=1G,tmp_requested=200G,tmpfree=200G #$ -q all.q # # echo "$HOST $tmp_requested $TMPDIR" dd if=/dev/zero of=$TMPDIR/dd.test bs=512M count=200 ------------------------------------------------------------------------------------------ $ cat dd_allq.o9 200G /scratch_local/9.1.all.q 200+0 records in 200+0 records out 107374182400 bytes (107 GB) copied, 73.2242 s, 1.5 GB/s So basically only prolog script has problem. Cheers, Derrick On Fri, Dec 7, 2018 at 9:52 AM Derrick Lin <klin...@gmail.com> wrote: > Hi all, > > We are switching to a cluster of CentOS7 with SGE 8.1.9 installed. > > We have a prolog script that does XFS disk space allocation according to > TMPDIR. > > However, the prolog script does not receive TMPDIR which should be created > by the scheduler. > > Other variables such as JOB_ID, PE_HOSTFILE are available though. > > We have been using the same script on the CentOS6 cluster with OGS/GE > 2011.11p1 without an issue. > > Thanks in advance. > > Cheers, > Derrick >
_______________________________________________ users mailing list users@gridengine.org https://gridengine.org/mailman/listinfo/users
