Hi AndyW and CarolinaDR (cc: DaveC):
(copied from an email I sent)
There is a missing file in the GL7 driving data on JASMIN for suite u-bb316:
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_vwind_1913_n96e.nc
AndyW is the owner/creator of this directory.
Can you transfer this file from somewhere?
Patrick
Hi Andy and Carolina:
In the meantime, I will try to run the suite starting in 1914 instead of 1860.
It probably won’t give the right results, but I can see if the suite works with the other changes I made.
Patrick
Hi Andy and Carolina:
Did you have a chance to look into this missing data yet?
Or is there someone else I should contact about this?
I think there is a 1940 data file missing, in addition to the 1913 data file.
Patrick
Hi Patrick,
File should have been updated. Let me know any issues.
Dr Andy Wiltshire
Hi again, Andy & Carolina:
Can you look into that missing 1940 data file sometime?
Thanks!
Patrick
Okay, done that one as well. It would be good to know if there are others.
Dr Andy Wiltshire
Thanks, Andy!
I will keep an eye out for for other missing files. I don’t know why they are missing. Maybe we need to get rid of group-write access for these files, if that is the current access.
Patrick
Hi Andy & Carolina:
I tried the run again of the u-bx723 suite which is derived from the u-bb316 GL7 suite, now starting from 1860 instead of 1914. It had previously failed in 1940.
It now gets to 1948 before failing due to missing data. This file is missing:
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_swdown_1948_n96e.nc
I guess there might be other files missing too.
Maybe it would be easier to just replace all the data all at once? Do you have an original copy of the data?
Patrick
Hi AndyW & Carolina:
Do you know where I can get an original copy of the driving data?
Then I could do this myself.
I looked on Monsoon in the path /data/d05/hadaw/HadGEM3-GA6/Ancils/GA7_Ancil/
but that appears to be a restricted Met Office path.
I found two more missing files, that makes at least 5 in total, and you’ve already replaced 2 of them. These are the two other missing files (there are likely more):
/gws/nopw/j04/jules/data/CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_tair_1966_n96e.nc
/gws/nopw/j04/jules/data/CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_lwdown_1963_n96e.nc
Patrick
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.
Someone noticed the ticket I had about this on the NCAS CMS Helpdesk, and they sent me a directory with another copy of the data.:
I just compared that directory to the official one I have been using.
I used this set of 3 commands to compare them:
dir1= set this to the other copy
dir2=/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e
for f1 in ${dir1}/*; do f2=“${dir2}/${f1#${dir1}/}”; [ ! -e “$f2” ] && printf ‘%s\n’ “$f2”; done
This is the result:
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_lwdown_1963_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_lwdown_1982_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_press_1972_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_press_2012_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_qair_2005_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_qair_2010_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_swdown_1948_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_swdown_1975_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_tair_1966_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_tair_1990_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_tair_1996_n96e.nc
/gws/nopw/j04/jules/data//CRU-NCEP-v7/CRU-NCEP-v7-n96e/cruncepv7_uwind_1999_n96e.nc
I tried to copy those files to the official database, but the folder belongs to AndyW, so I don’t have permission to update the files.
Patrick
Hi DaveC:
Since we now have a full CRU-NCEP-v7 dataset available, can you move this directory from:
/gws/nopw/j04/jules/data/CRU-NCEP-v7/
to:
/gws/nopw/j04/jules/data/CRU-NCEP-v7_oldCorrupt
?
And can you then move this directory:
/gws/nopw/j04/jules/data/CRU-NCEP-v7_copy2/
to:
/gws/nopw/j04/jules/data/CRU-NCEP-v7/
You can recursively change the owner of the files in the last directory to yourself, if you want.
I have made those files in that directory read-only, possibly to help to avoid files going AWOL like before.
I have tested the suite u-da046 with this copy2 of the data. But there are other people’s GL7 N96 suites that might get tripped up by missing files in the original copy of this CRU-NCEP-v7 dataset.
Patrick
If you have access to that directory, why do you want me to copy the directories? You have permission yourself, no?
Hi DaveC
I want to replace the corrupt copy that has missing files. I don’t have permission to do that.
Patrick
I’ve done that.
Do you want to keep the _oldCorrupt directory?
I’ve not looked into this, just done your commands. If you’d like to have access to the julesadmin unix group, I can get you that? You don’t have to if you don’t want to.
Thanks, Dave!
That’s perfect.
I think we should keep the prior copy of the data for awhile, in case there are problems.
I don’t think there are enough times that I would need julesadmin privilege, so I will hold off on that for now.
Patrick