Opened 10 months ago

Closed 9 months ago

#2749 closed help (fixed)

Difficulty migrating to xcslc0

Reported by: anmcr Owned by: ros
Component: Monsoon Keywords:
Cc: Platform: Monsoon2
UM Version: 11.1

Description

Hello,

I'm having issues running the rose editor on xcslc0. I followed the advice given at: https://code.metoffice.gov.uk/trac/home/wiki/AuthenticationCaching#Monsoon. I also followed: https://collab.metoffice.gov.uk/twiki/bin/view/Support/RetirementOfRoseCylcVMs.

However, when I try to run 'rose edit' or edit my jobs through 'rosie go' then I get the following error (which is similar to ticket #2732, but not alleviated by using 'module unload python'):

xcslc0:/home/d01/amworr$ module unload python
xcslc0:/home/d01/amworr$ cd roses
xcslc0:/home/d01/amworr/roses/u-be146$ rose edit&
[1] 138943
xcslc0:/home/d01/amworr/roses/u-be146$ [FAIL] Could not load macro /home/d01/amworr/roses/u-be146/meta/lib/python/macros/nesting_suite.py: Traceback (most recent call last):
[FAIL] File "/common/fcm/rose-2018.06.0/lib/python/rose/macro.py", line 729, in load_meta_macro_modules
[FAIL] modules.append(imp.load_source(as_name, meta_file))
[FAIL] File "/home/d01/amworr/roses/u-be146/meta/lib/python/macros/nesting_suite.py", line 7, in <module>
[FAIL] import pytz
[FAIL] ImportError?: No module named pytz

I would be grateful for any help.

Thanks,

Andrew

Attachments (1)

for_ros.JPG (62.3 KB) - added by anmcr 10 months ago.

Download all attachments as: .zip

Change History (15)

comment:1 Changed 10 months ago by ros

  • Owner changed from um_support to ros
  • Status changed from new to accepted

Hi Andrew,

It's because this suite has a macro included that uses a python module that isn't available. I will let the Monsoon team know, but until then you will still be able to edit and run the suite ok on the login nodes, just won't be able to run that particular macro. If you do need to run that macro you should probably continue using exvmsrose for the moment.

I'll let you know when this has been fixed or I have a workaround.

Regards,
Ros.

comment:2 Changed 10 months ago by anmcr

Hi Ros,

Thanks for looking into thia. I've been continuing to use evxsmrose, but I'm aware that is switched off on 12 Feb. Will the xcslc servers be available to me by this date, so that I continue my model runs. I'm afraid that I have quite a tight timeframe to get them completed.

Thanks,

Andrew

comment:3 Changed 10 months ago by ros

Hi Andrew,

I have just heard from the Met Office that the missing pytz python module will be installed in the next hour or so.

Could you try firing up rose edit on xcslc* later this afternoon or tomorrow morning and let me know if that fixes the problem.

Thanks.
Ros.

comment:4 Changed 10 months ago by anmcr

Hi Ros,
Thanks for the update.
Yes, I will do this later this afternoon.
Best wishes,
Andrew

comment:5 Changed 10 months ago by anmcr

Hi Ros,

I was able to run 'rose edit' successfully on xcslc.

Thanks for your help.

Andrew

comment:6 Changed 10 months ago by anmcr

Hi Ros,

I'm getting a lot of 'submit-failed' issues, which require me to trigger the run again. I didn't have any problems like this when I was running on exvmsrose. Is this a monsoon wide issue, or can I do something about it?

Thanks,

Andrew

comment:7 Changed 10 months ago by ros

Hi Andrew,

Without seeing what the error message is it's very difficult to say. I had a quick look in a couple of your suites but couldn't find any of the failures. It's possible that you are being hit by the qsub bug, but can't say for sure. Full details on this can be found here: https://collab.metoffice.gov.uk/twiki/bin/view/Support/Monsoon2BugInQsub

If this is not the problem please tell me when it next happens; give me the suite id and the name of the task that has failed and I'll take a look.

Regards,
Ros.

comment:8 Changed 10 months ago by anmcr

Hi Ros,

I don't understand this, as I am sure that I checked properly last week, but I am getting more 'No module named pytz' errors on xcslc*. See below. Could you please have another look at this.

Many thanks for your help,

Andrew

xcslc0:/home/d01/amworr/roses/u-bf575$ [FAIL] Could not load macro /home/d01/amworr/roses/u-bf575/meta/lib/python/macros/nesting_suite.py: Traceback (most recent call last):
[FAIL] File "/common/fcm/rose-2019.01.0/lib/python/rose/macro.py", line 747, in load_meta_macro_modules
[FAIL] modules.append(imp.load_source(as_name, meta_file))
[FAIL] File "/home/d01/amworr/roses/u-bf575/meta/lib/python/macros/nesting_suite.py", line 7, in <module>
[FAIL] import pytz
[FAIL] ImportError?: No module named pytz

comment:9 Changed 10 months ago by anmcr

Hi Ros,
The 'pytz' module problem must have been resolved last week, as I was editing my jobs on xcslc*.
Best wishes,
Andrew

comment:10 Changed 10 months ago by ros

Hi Andrew,

The Met Office are aware. Unfortunately when the patching was done the other day the module fix hadn't made it into the image so it vanished. This will hopefully be fixed tomorrow.

Regards,
Ros.

Changed 10 months ago by anmcr

comment:11 Changed 10 months ago by anmcr

Dear Ros,

I am able to use 'rose edit' now. However, I now recieve a 'submit failed' error under 'INSTALL_COLD'. See attachment. A colleague flixed this in another version of the UM by setting 'host=localhost' in the '../site/MONsoon.rc'. But I was not able to find a similar file for my setup. The job id is u-bf574. Could you please advise.

Many thanks,

Andrew

comment:12 Changed 10 months ago by ros

Hi Andrew,

I'm not sure why it won't work submitting to postproc I shall report this, however in the meantime please change the host to be localhost. This is in file site/monsoon-cray-xc40/suite-adds.rc.

Near the top of the file change:

IDL_SERVER = "postproc"

to be

IDL_SERVER = "localhost"

Similarly (on first line) change:

HPC_HOST = "xcs-c"

to be

HPC_HOST = "localhost"

Cheers,
Ros.

comment:13 Changed 10 months ago by anmcr

Hi Ros,

That worked. Many thanks for all your prompt help. It is very much appreciated. Please close this ticket.

Best wishes,

Andrew

comment:14 Changed 9 months ago by ros

  • Resolution set to fixed
  • Status changed from accepted to closed
Note: See TracTickets for help on using tickets.