[GE users] when SGE emails don't work on Apple OS X (problem described & workaround)

craffi dag at sonsorol.org
Thu Feb 19 16:00:08 GMT 2009

This is the 2nd time I've seen (and been frustrated by) this issue,  
spanning PPC and Intel architecture Apple boxes running different  
versions of SGE 6.0 or 6.1

If you encounter these symptoms:

> (1) /usr/bin/mail works perfectly from the command line
> (2) /usr/bin/mail configured as the SGE mailer produces no email or  
> useful log entries
> (3) substituting a wrapper with extra logging also produces no logs  
> or email

*And* your only clue is this sort of output in the execd spool logs:

>> 09/10/2008 16:22:07|execd|xxx-fs01|E|mailer had timeout - killing
>> 09/10/2008 16:22:07|execd|xxx-fs01|E|mailer exited with exit status  
>> = 1
>> 09/10/2008 16:22:19|execd|xxx-fs01|E|mailer had timeout - killing
>> 09/10/2008 16:22:19|execd|xxx-fs01|E|mailer exited with exit status  
>> = 1

... then the solution seems to involve explicitly overwriting  
DYLD_LIBRARY_PATH for the SGE mailer script:


Make a wrapper script like this:

> #!/bin/sh
> export DYLD_LIBRARY_PATH=/usr/lib
> /usr/bin/mail -s "$2" $3

Save the wrapper somewhere where all your hosts can get at it.

And change the mailer parameter in the main SGE configuration (qconf - 
mconf) as well as for all the hosts that have local mailer settings  
(qconf -mconf <hostname>) to point to the new mail wrapper.

You might also want to increase the verbosity of the OS X mailer  
program which by default is pretty quiet:

# serveradmin settings mail:postfix:log_level="info"

I'm posting this here as google bait in case others have seen the same  
issue. I don't think I posted this before, if so apologies for the  
double post.



To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].

More information about the gridengine-users mailing list