# Running HadGEM3-GC3.1-LL from Puma

**URL:** <https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345>\
**Category:** Rose/Cylc and FCM\
**Tags:** ARCHER2\
**Created:** [7 December 2021 16:03 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345 "2021-12-07T16:03:14Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![racheldiamond](https://avatars.discourse-cdn.com/v4/letter/r/eb9ed0/32.png) [@racheldiamond](https://cms-helpdesk.ncas.ac.uk/u/racheldiamond)\
**Post date:** [7 December 2021 16:03 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345/1 "2021-12-07T16:03:14Z")

</div>

Dear NCAS CMS Helpdesk,

I am trying to run my first job on ARCHER2 from PUMA. My suite is u-ck049 - this is a copy of u-as037 (which I believe can be used for an experiment like CMIP6 piControl, but is modified to run on ARCHER2). I have made minor modifications to the suite and also have changed the runtime from 500 years to 1 month.

However when I try to run u-ck049, I get the following error in fcm\_make\_ocean: 'No such file or directory: ‘/export/puma/data-01/training//fcm\_make\_ocean’. I think this is because I originally set up my account for the UM training course, and my ‘.profile’ file includes the line: ‘export SCRATCH=/export/puma/data-01/training/$USER’.

Is there something I can change in the ‘.profile’ file so that I can run suites on ARCHER2 from PUMA? Or do I need to set up an entirely new login account?

Many thanks.

Best wishes,

Rachel

---

<div class="post-metadata">

**Author:** ![RosalynHatcher](https://dub1.discourse-cdn.com/flex013/user_avatar/cms-helpdesk.ncas.ac.uk/rosalynhatcher/32/11_2.png) [@RosalynHatcher](https://cms-helpdesk.ncas.ac.uk/u/RosalynHatcher)\
**Post date:** [7 December 2021 16:51 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345/2 "2021-12-07T16:51:18Z")

</div>

Hi Rachel,

Are you trying to run on the full 23cabinet system? If so we haven’t yet updated this suite to run on the full system - it’s on my to-do list for this week.

Andy has just created you a SCRATCH directory on PUMA at: `/export/puma/data-01/radiam24`. Please update the path in your `~/.profile`.

Regards,  
Ros.

---

<div class="post-metadata">

**Author:** ![racheldiamond](https://avatars.discourse-cdn.com/v4/letter/r/eb9ed0/32.png) [@racheldiamond](https://cms-helpdesk.ncas.ac.uk/u/racheldiamond)\
**Post date:** [8 December 2021 11:25 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345/3 "2021-12-08T11:25:04Z")

</div>

Hi Ros,

Thanks so much, I’ve done that. No, I’ve been trying to run on the 4-cabinet system, should this work?

The suite now starts running but fails at the ‘coupled’ task; processors 0 to 5 all return an error related to the IO server. I’m confused by this as I haven’t changed any IO server settings from u-as037; the only change I’ve made is removing the ‘validate\_suite\_info’ task as in [http://cms.ncas.ac.uk/ticket/3404?cversion=0&cnum\_hist=2](http://cms.ncas.ac.uk/ticket/3404?cversion=0&cnum_hist=2). Could you please let me know what to do about this, or suggest a different suite I could use to simulate the PI using HadGEM3-GC3.1-LL on ARCHER2? Many thanks.

The errors returned look like:

Error code: 99  
? Error from routine: IOS\_INIT:IOS\_RUN  
? Error message: IO Server is not active  
? Error from processor: 3  
? Error number: 0  
???

[47] exceptions: An non-exception application exit occured.  
[47] exceptions: whilst in a serial region  
[47] exceptions: Task had pid=33264 on host nid001007  
[47] exceptions: Program is “/lus/cls01095/work/n02/n02//cylc-run/u-ck049/work/18500101T0000Z/coupled/./atmos.exe”  
Warning in umPrintMgr: umPrintExceptionHandler : Handler Invoked

---

<div class="post-metadata">

**Author:** ![grenville](https://dub1.discourse-cdn.com/flex013/user_avatar/cms-helpdesk.ncas.ac.uk/grenville/32/29_2.png) [@grenville](https://cms-helpdesk.ncas.ac.uk/u/grenville)\
**Post date:** [8 December 2021 11:44 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345/4 "2021-12-08T11:44:02Z")

</div>

Hi Rachel

Please try setting

`ios_num_threads = 2`

Grenville

---

<div class="post-metadata">

**Author:** ![racheldiamond](https://avatars.discourse-cdn.com/v4/letter/r/eb9ed0/32.png) [@racheldiamond](https://cms-helpdesk.ncas.ac.uk/u/racheldiamond)\
**Post date:** [8 December 2021 14:36 UTC](https://cms-helpdesk.ncas.ac.uk/t/running-hadgem3-gc3-1-ll-from-puma/345/5 "2021-12-08T14:36:52Z")

</div>

Hi Grenville,

Thanks for the suggestion, ios\_num\_threads was originally a latent variable so I added it, set it to 2 as you suggested, and ran the suite again.  
However I am still getting the same error, is there anything else I could change?

Best wishes,

Rachel