# Restart simulations

**URL:** <https://forum.ecmwf.int/t/restart-simulations/12120>\
**Category:** OpenIFS\
**Tags:** openifs-configuration-and-namelists, openifs-40r1, openifs-long-runs\
**Created:** [15 November 2016 11:27 UTC](https://forum.ecmwf.int/t/restart-simulations/12120 "2016-11-15T11:27:48Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:27 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/1 "2016-11-15T11:27:48Z")

</div>

I have run a long forecast and want to re-start it. I have the srf files and the rcf file. When I try to restart the forecast (without making any changes to fort.4 or rcf) nothing seems to happen - i.e. no more timesteps are calculated and thus no additional output is produced. The simulation does not fail though. Do I need to change something in either fort.4 or rcf? I have attached the NODE file.

[NODE.001\_01.gz](https://forum.ecmwf.int/uploads/short-url/lFfeDEXcTaasqM7Q4axERtRvlZ.gz) (124.5 KB)

---

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:28 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/2 "2016-11-15T11:28:05Z")

</div>

Hi Victoria,

When you run the restart do you increase the value of NSTOP? Otherwise the model will read the restart files and not progress because it thinks all the timesteps are complete.

&nbsp; Glenn

---

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:28 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/3 "2016-11-15T11:28:27Z")

</div>

Thanks - I'm fairly sure this is the problem. I did change NSTOP but assumed that the start of the restart was time 0 again. Clearly I need to add the restart forecast length to NSTOP at the end of the first forecast. I'll let you know if it works - the job is currently in a queue.

Victoria

---

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:33 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/4 "2016-11-15T11:33:17Z")

</div>

When I have set NSTOP to an appropriate value, I get this error in my restart forecast:

GRIB\_API ERROR : grib\_set: missingValue Invalid grib id

Any idea what the problem is? Interesting the simulation did not fail - it just seemed to hang and not produce any output.&nbsp;

Victoria

---

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:33 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/5 "2016-11-15T11:33:43Z")

</div>

Hi Victoria,&nbsp; I'm not sure what's causing that. I don't recall seeing that before. Can you send the NODE file?

Thanks,&nbsp; Glenn

---

<div class="post-metadata">

**Author:** ![Glenn\_Carver](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/g/e495f1/32.png) [@Glenn\_Carver](https://forum.ecmwf.int/u/Glenn_Carver)\
**Post date:** [15 November 2016 11:34 UTC](https://forum.ecmwf.int/t/restart-simulations/12120/6 "2016-11-15T11:34:35Z")

</div>

Here is the Node file. I'm not sure it is helpful as the job does not fail - it just hangs without producing output. In this case I've killed the job so there is no trace back

Victoria

[NODE.001\_01.gz](https://forum.ecmwf.int/uploads/short-url/lFfeDEXcTaasqM7Q4axERtRvlZ.gz) (124.5 KB)
