This is a known issue, and we are currently working on a fix. The best
way to address this problem for now is to set
EXECUTE_LOGIN_IS_DEDICATED = True
in the config file. This will not cure the problem entirely, but it
makes it much less likely to occur. The only concern with setting this
parameter is that you must be sure that your jobs are running under a
dedicated account. By default, they will run as condor-reuse-vm1 (or
2..n depending on how many VMs the machine has), which is obviously a
dedicated account and is safe to use with this parameter.
If you happen to be using the VMx_USER feature in the config file, you
should be aware that any process running under the specified account
will be treated as part of the job, which could lead to unexpected results.
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}