Hi Ros,
A few months ago I added back in (see my job xpdb-a) the GLOMAP-aerosol to the v8.4 GA4 UM-UKCA job (Release Job 4.0) which seemed to be missing from the xoxta job that NCAS-CMS ported to ARCHER-2,
And then about a month or so ago, I completed the upgrade steps for specific jobs that upgrade from the “GLOMAP-mode v7newprim” codebase within the RJ4.0 release job, to then the standard UM-UKCA jobs with GLOMAP v8.1 from Yoshioka et al., 2019, JAMES (see my job xpdb-y – copy of Masaru’s base job teaf-w).
And then I furthermore completed the upgrade-steps to the main GLOMAP v8.2 codebase (see my job xphr-f – copy of Sandip’s SMURPHS base job xnbe-b) that was applied with “v3 of UM-UKCA” for the range of strat-aerosol/volc-aerosol papers in 2017-2021 (e.g. in quiescent conditions – Brooke et al., 2017, JGR; for the Tambora-ISA experiment – Marshall et al., 2018, ACP; Clyne et al., 2021, ACP – for the major-volcanic forcing PPE – Marshall et al., 2019, JGR; and for the SMURPHS/ACSIS volcanic forcing datasets Dhomse et al., 2020, ACP; Antuna-Marrero et al., 2021 EESD; Feng et al., zenodo-published waveband-mapped datasets ).
This was working fine in early February, and I set up and ran a set of GA4 UM-UKCA interactive stratospheric aerosol simulations for the Hunga-Tonga aerosol cloud (see UMUI experiment xphy)
Anyway – the reason I’m raising this post – as well as to flag up the availability of these standard GA4 UM-UKCA trop-aerosol & strat-aerosol “standard jobs” (at GA4 UM-UKCA in v8.4 of the UM) – is re: the problem with PUMA and the temporary switch to “pumatest”.
I was in touch with Andy Heaps last week (see emails below) with the initial submit of the standard “strat-trop v3” GA4 UM-UKCA job xphy-s (with GLOMAP v8.2) was failing – due to not being able to access an “allow_aerosols” user STASHmaster file in Mohit Dalvi’s home directory on PUMA.
It turned out that at that time, Mohit’s mdalvi directory had not yet been re-instated on to the pumatest machine.
After my email below asking for Mohit’s directory to be re-instated, that was then subsequently done, and then my submission of the standard xphy-s job worked fine.
However, I tried again on Friday afternoon, and it turns out that the job also uses 2 hand-edits and 3 user-STASHmaster files within Dan Partridge’s user directory dan2012
I managed to find the 2 hand-edits and 2 of the 3 user-STASHmaster files, and copied those over to my directory and the job then progressed a bit further – but still there is a 3rd user-STASHmaster from Dan’s user directory that I couldn’t find a copy of (kohler_dp_8.4)
***1) Please can you also re-instate Dan Partridge’s user-directory dan2012 on the temporary pumatest system? (as you already did for Mohit’s mdalvi directory) ***
Also – I did try just submitting without that user-STASHmaster – and the job fails with not finding the ukca directory for a series of hand-edits used in that standard job xphy-s/xnbe-b:
~ukca/hand_edits/VN8.4/config_new_diags_extra.ed
~ukca/hand_edits/VN8.4/sect35_on.ed
~ukca/hand_edits/VN8.4/config_strattrop_jpbk.ed
~ukca/hand_edits/VN8.4/CFC-114_not_in_Rad.ed
~ukca/hand_edits/VN8.4/UKCA_useUMvals.ed
2) Please can you also re-instate the ukca user-directory on pumatest
I set these to be N rather than Y – and that gets round the problem with the job xphy-s now completing the “UMUI process stage” OK.
However, when I click submit, there is then a more fundamental problem – a problem during the FCM build of the branches as the code progresses:
MAIN_SCR: Calling Extract …
Extracting UMATMOS base repository…
UMATMOS base repository extract failed
See extract output file /home/gmann/um/um_extracts/xphys/baserepos/UMATMOS/ext.out
MAIN_SCR: Extract failed
MAIN_SCR stopped with return code 25
I did check whether the FCM commands were working OK when I first ssh’d in to pumatest (see my emails below with Andy H) and the fcm status commands seemed to be working fine, in communicating with the PUMA FCM repository etc.
However – it looks like the FCM build process triggered from the UMUI is not able to progress at the moment – probably just missing a file or some other file-path problem?
Also – I noticed that PUMA UM Trac web-system doesn’t seem to be working.
When I point a browser to https://puma.nerc.ac.uk it just hangs.
I’ve attached a screen-shot of the
From: Graham Mann
Date: Wednesday, 16 February 2022 at 17:12
To: Andy Heaps
Subject: Re: pumatest
Hi Andy,
OK, great – thanks for your help with this.
I can already be getting on with the extra FCM-commits for the planned code-changes
for the shorter-duration (more highly concentrated) volcanic emission of SO2.
The Hunga-Tonga eruption emitted 0.2-0.4Tg of SO2 in just 20 minutes, compared to Pinatubo’s 14-23Tg of SO2 that was emitted over 9 hours — H-T was very explosive and also very concentrated chemistry!!
Cheers
Graham
From: Andy Heaps
Date: Wednesday, 16 February 2022 at 17:09
To: Graham Mann
Subject: Re: pumatest
Hi Graham,
I think Ros is going to do this tomorrow. Time to get back on strike!
Cheers
Andy
On 16/02/2022 16:54, Graham Mann wrote:
Hi Andy,
I’ve logged in to the pumatest machine from remote-access.leeds.ac.uk and can open the
UMUI OK (see the 2nd of the two attached screen-shot PNG files).
And I can also see that FCM commands seem to be working OK from pumatest
However, when I tried to submit a test UMUI job (my recent Hunga-Tonga control job xphy-s)
it fails when trying to access a file within Mohit Dalvi’s directory on PUMA.
This worked fine about 2 weeks ago, and I’m assuming then it’s simply a case of you not having
restored his user directory on the network visible from the pumatest machine.
See the 1st PNG file gives the error message it can’t access the file “allow_aerosols.stash”, which is a user STASHmaster file within Mohit Dalvi’s directory ~mdalvi/umui_jobs/prestash/vn8.1 directory.
Please can you restore Mohit Dalvi’s directory mdalvi on this temporary system?
Thanks a lot
Best regards,
Cheers
Graham
PS I’m supposed to be on strike too!
On 16/02/2022, 16:32, “Graham Mann” wrote:
Hi Andy,
OK -- thanks for this.
I'll try this now -- logging in from the new remote-access.leeds.ac.uk server.
Cheers
Graham
On 16/02/2022, 15:40, "Andy Heaps" wrote:
Hi Graham,
PUMA has some issues so moving to a temporary server and we've restored
your home directory from the backups on the evening of the 14th
February. Your ssh key should work as normal just that the server name
has changed to pumatest.nerc.ac.uk. I have reset your password to be
**** deleted for NCAS-CMS Helpdesk post **** Please change this when you login.
Note:
Your shell has changed to the bash shell
The cylc-run directory isn't backed up so won't be present in your home
directory
A further email will come in due course on how to restart any running suites
I intend to go back on strike again tomorrow so if you have any queries
after the close of play today should go to the CMS help desk.
Cheers
Andy
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Andy Heaps
National Centre for Atmospheric Science (NCAS)
Room 118
Harry Pitt Building,
Reading University,
Earley Gate,
PO Box 243,
Reading RG6 6ET
U.K.
tel: 0118 378 6421
fax: 0118 378 8316
e-mail: andy.heaps@ncas.ac.uk
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
–
Andy Heaps
National Centre for Atmospheric Science (NCAS)
Room 118
Harry Pitt Building,
Reading University,
Earley Gate,
PO Box 243,
Reading RG6 6ET
U.K.
tel: 0118 378 6421
fax: 0118 378 8316
e-mail: andy.heaps@ncas.ac.uk
