> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
>
> For your delight and delectation:
<snip>
>
> ANALY_OX: Good, but again running short on local disk space: "Error
> details: pilot: Too little space left on local disk to run job:
> 1555456 kB (need > 2097152 kB)"
>
Not sure what's happening here; we've got 500G disks in the nodes, so
we should have the full 50G per job slot actually available;
furthermore,
I've just had a look at all the nodes as of now and the most full one
has 320Gb of free space.
It seems to me that there are a couple of possibilities here; either
there's a problem with the check (doesn't sound likely) or we've got
jobs that are filling the WN disks, then cleaning up after themselves,
and I'm not sure how best to spot that. We do have nagios alerts for
this, but it only polls every few minutes, so it could miss a short-
lived problem.
Ewan
|