Re: [HTCondor-users] best way to use cached data

[Matthew Farrellee <matt@xxxxxxxxxx>]

On 01/12/2013 12:21 PM, Dimitri Maziuk wrote:
> On 1/12/2013 6:26 AM, Matthew Farrellee wrote:
>
> > I've had success with startd cron for advertising the contents of a
> > cache, and highly recommend it over configuration changes.
>
> As I understand startd cron, you configure your nodes to periodically
> run a script that publishes custom attributes, and the way a regular

Yes, where you == administrator.


> user would publish their custom attributes is by modifying the script.
> (Or am I missing something?)

No, the administrator needs to control the script. It should: gather | 
sanitize | publish. gather could be as simple as ls /thecache.


> The script runs as condor (root) user, so security-wise this is worse
> than letting them 'sudo condor-reconfig' as now they can run anything as
> condor.

Security-wise giving the job owner the ability to change configuration 
on a node is equivalent to giving them a setuid-root script to run.


> Either way, my point was that doesn't work when you're shipping
> (flocking, gliding) jobs off-site and have no control over execute nodes
> whatsoever.

If you are crossing administrative domains, neither solution will work 
without cross-organizational agreement. I expect you'll have a simpler 
time asking for a startd cron to be run than allowing jobs to alter node 
configuration.


Best,


matt