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

Re: [HTCondor-users] Question about flocking



As I understand it, the user for a Flocked job is effectively nobody. Would your test succeed silently if the user isn't in the directory in the first place? I haven't used Flocking before, I just read the 1996 paper and the Flocking documentation yesterday.

On 1/29/15 10:40 AM, Ian Cottam wrote:
On our Pool we have a single submit node. There we have proxy scripts for,
e.g., condor_submit that do extra checks.
In one case -- the absence of a request_memory line -- we don't run the
job.

Recently, I have been testing Personal Condor and flocking to our main
Pool.
I fully appreciate that the above proxy scripts will never be invoked
(obviously). However, we have a job hook on our compute nodes that checks
for a classad, call it, HAS_APP, and then looks up in an active directory
if the user has permission to run APP. This simple check works fine when I
submit from our main Pool's submitter, but seems to be being ignored if I
flock to the Pool. That is, any user can run APP.

Is this behaviour to be expected and if so why?
Thanks
-Ian


--
Gary