 |
 |
On 21 May 2012, at 11:34, emyr.james wrote: > Dear all, > > I'm still trying to get the Cream-CE working here at sussex. > Here's what I'm doing... > The error in the StandardError file is... > > [root@node201 dteam154]# cat err_cream_782051410_StandardError > connect: No route to host at -e line 23. > connect: No route to host at -e line 23. > connect: No route to host at -e line 23. > > Does anyone have any ideas on where to go from here ? Line 23 in which > file is it complaining about ? This is a problem with the file stage out, or possibly stage in. The line 23 is (almost) certainly the qsub'd script that was sent into SGE (from CREAM). if you can capture that script, check what line 23 says (-e is the place to put stderr, which is nomally specificed in (rcp|scp) language, as user@host:path/to/file). It's could be something like the CREAM host is putting in the internal name for itself, which is not in the hosts file on the WN's (whereas the external name is in the DNS). Eitherway, once you see the -e directive, you'll probably be able to work out what's wrong pretty quick. It _might_ be other places, but in my experience of breaking the CREAM/BLAH/LRMS links, that looks to me like the LRMS complaining about the job config it got from CREAM.