[Ncep.list.nems.announce] NEMS commit to restore support for /gpfs/hps/ptmp
Samuel Trahan - NOAA Affiliate
samuel.trahan at noaa.gov
Thu Jul 20 17:20:30 UTC 2017
This change is committed to all three supported apps (NEMSGSM, WW3-GSM, and
NEMSfv3gfs). Now that the compset/regtest data transfers are complete, and
this bug is fixed, all should be well again.
On Thu, Jul 20, 2017 at 11:17 AM, Samuel Trahan - NOAA Affiliate <
samuel.trahan at noaa.gov> wrote:
> Hi all,
> Recall I modified the NEMSCompsetRun last week so the filesystem space
> auto-detector will scan the new /gpfs/hps/[sp]tmp filesystems on Luna.
> I tried to retain backward compatibility to the old filesystem structure,
> but had no way to test that. The unimaginable has happened: I broke
> backward compatibility. I'm going to commit a change shortly, around 12:30
> PM eastern, that fixes this bug. It has been tested with the only app that
> supports WCOSS Cray: NEMSfv3gfs.
> This also hides a confusing message you'll see when the script realizes
> the new filesystems don't exist on Surge.
> Index: tests/rtgen
> --- tests/rtgen (revision 95891)
> +++ tests/rtgen (working copy)
> @@ -688,7 +688,7 @@
> [ '/ptmpp2', '-j', 'ptmp-p2', 'gpfs-'+tg+'p2'] ]
> max_area='/ptmpp2' # default on failure
> else: # assume cray (pex=0)
> - ptmps=[ [ '/gpfs/hps', '-j', 'hps-ptmp', 'hps'],
> + ptmps=[ [ '/gpfs/hps/ptmp', '-j', 'hps-ptmp', 'hps'],
> [ '/gpfs/hps/stmp', '-j', 'hps-stmp', 'hps'],
> [ '/gpfs/hps2/ptmp', '-j', 'hps2-ptmp', 'hps2'],
> [ '/gpfs/hps2/stmp', '-j', 'hps2-stmp', 'hps2'],
> @@ -700,6 +700,8 @@
> for ptmp in ptmps:
> + if not os.path.exists(ptmp):
> + jlogger.info('%s: does not exist'%(ptmp,))
> args=['/usr/lpp/mmfs/bin/mmlsquota', '--block-size', '1T' ]
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Ncep.list.nems.announce