[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Lost address file



When using condor_cold_start it is best to place the -localdir on the local filesystem or AFS rather than NFS. Using the local file system will reduce NFS idiosyncrasies and increase performance.

_joe

Ryan Garver wrote:
I'm having the same problems with the same characteristics with 6.7.18. Any suggestions?

Ryan Garver wrote:
Ryan Garver wrote:
I'm having a strange problem with my address files that are being put
down by the master and schedd daemons.  I have all of the *_ADDRESS_FILE
settings filled out.  The strange thing that I'm experiencing is that
the .master_address file is created, and then disappears.  It also
appears that my .schedd_address never is created.  Every thing seems to
be performing properly and if I set the -name parameter for condor_q I
can see the schedd's queue correctly.  This is annoying from the users
perspective (requiring the -name parameter) and also when I shut things
down I get errors because the schedd can't be found (BTW this is a
personal condor install).  What could cause this, and how might I fix
it?  Thanks

Just to give you some more information on this.  I am using the
cold_start tools for this.  My main config file looks like:

<<< Snip

MASTER_ADDRESS_FILE = $(EXECUTE)/.master_address
STARTD_ADDRESS_FILE = $(EXECUTE)/.startd_address
SCHEDD_ADDRESS_FILE = $(EXECUTE)/.schedd_address
CREDD_ADDRESS_FILE  = $(EXECUTE)/.credd_address

<<<

and EXECUTE is set in condor_config.local_install as:

EXECUTE = /cs/sandbox/student/rgarver/condor_install/conf.snoopy/execute

This directory does exist and has in the past yielded a .master_address
and .schedd_address and in fact does produce one when I start a single
personal condor locally using the same sets of files and binaries.  I am
executing this over ssh with all of files for condor pre-staged, that
seems to be the only difference.  As I said above, the .master_address
at least (.schedd_address may also) makes a brief appearance and then
disappears.  This is not mentioned at all in any of the logs.  The whole
system continues working properly however.  Without these files shutdown
becomes very messy.
I would appreciate any suggestions on how I might continue to debug this
and resolve the issue, Thanks.