Atlas-connect-l mailing list

[Christopher Meyer <chris.meyer AT cern.ch>]

Hi Lincoln,

No problem, thanks for the fast action!

I do have a question about the best place to store different things. At UCT3 I had all the condor output logs sent to /share/t3data to avoid slamming the home disk. Should I do something similar from ATLAS connect? I noticed a link to stash/ in my home directory, but it points at a non-existent path.

Thanks!

Chris

On Wed, Feb 17, 2016 at 7:12 PM Lincoln Bryant <lincolnb AT hep.uchicago.edu> wrote:

Hi Chris,

I've freed up about 500GB and we're asking users to migrate data to
appropriate places.

Sorry for the inconvenience!

Thanks,
Lincoln

On Wed, 17 Feb 2016, Christopher Meyer wrote:

> Dear Experts,
>
> Last night while running, my condor jobs went into recovery mode, and
> eventually aborted due to mis-matched job Ids in various log files. Looking
> back it bit, it seems to have all started because the home disk is full.
> Luckily submitting again using the same input dag file seems to have found
> the recovery file, and things picked up where they left off.
>
> The various users added together seem to come in at under 0.5 TB (although
> I'm leading the way with 200 GB or so). All 7.3 TB of this disk are used
> though, based on a quick df -h.
>
> Could you please take a look?
>
> Thanks!
> Chris
>