Cycling nesting suite fails after two runs

[tstein]

Hi,

I want to run the Nesting Suite as 36-hour forecasts for 30 separate days (suite u-ao745).
It manages to run the first two days perfectly, but for the third day I get the following error:

[job-submit cmd] cylc jobs-submit —host=login.archer.ac.uk —user=thmstein —remote-mode — '$HOME/cylc-run/u-ao745/log/job' 20161102T1800Z/glm_um_recon1/01
[job-submit ret_code] 38
[job-submit out] 2018-01-10T17:25:28Z|20161102T1800Z/glm_um_recon1/01|38|None
(thmstein@…) 2018-01-10T17:25:28Z [STDERR] qsub: would exceed queue generic's per-user limit

The CRUN length for my suite is 3 hours, so it probably submits 12 jobs per forecast day.
I also have "Build new executable" switched on, so that's a task for each forecast day - I probably do not need this?
I also have "Create new ancillaries", but I believe that is necessary for each forecast day, as it will be nested inside a different global analysis run.

Is there a set up that I should consider that would allow my entire run to succeed without hitting my user limit for submitting jobs to the queue?

Cheers,
Thorwald

[grenville]

Cylc can do this for you see

​https://cylc.github.io/cylc/html/multi/cug-htmlse12.html#x13-20800012.11

add to suite.rc

[scheduling]

[[ queues]]

[[[serial]]]

limit = 3 (for example)

Grenville

[grenville]

Thorwald

I forgot to add a members = entry —- that maybe tricky for the nesting suite

Alternatively, you could just set the default

[[ queues]]

[[[default]]]

limit = 2 (for example)