Am 04.04.2012 um 09:12 schrieb Ursula Winkler: > Reuti wrote: >> >>> Ursula Winkler wrote: >>> Well, obviously the SGE doesn't find ANY path - the epilog routine is >>> also not found. Has anybody seen such an behaviour before? >>> >> >> Can you submit a simple job with: >> >> df -h >> >> The job script is executed with the same environment like the prolog,... >> Inside the jobscript you can access /bin/true and execute it? >> > "df -h" within the job-file is executed and shows also that all shares are > there. So the path is found within a job. > > I have a suspicion: when I compiled the SGE I did it with Berkeleydb > "db4-devel" by CentOS and did not install the version from > http://www.oracle.com/technetwork/database/berkeleydb/overview . Maybe this > is the problem? > What is preferred: spooling with berkeley or classic? I never tried classic > spooling so I cannot say if there any > disadvantages. Is it better to make a new install of the master with classic > spooling?
This is also some kind of personal taste. Some prefer classic spooling as you can check all the information of a job as they are just stored as text files. And it even handles a large number of nodes before it gets performance problems. Maybe Chris can make a statement about it as he runs large clusters with classic spooling IIRC. You tried a file in / and it's working then, e.g. copy "/bin/true" to "/true"? -- Reuti > > Ursula > _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
