# StreamTracer loose tracers without Reasonfortermination
**URL:** https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247
**Category:** ParaView Support
**Tags:** vtk
**Created:** [May 4, 2020, 3:29pm UTC](https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247 "2020-05-04T15:29:41Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@Jo\_Hey](https://discourse.paraview.org/u/Jo_Hey)
#### Post date: [May 4, 2020, 3:29pm UTC](https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247/1 "2020-05-04T15:29:41Z")
Hi,
I am using the streamtracer vtk filter since a while and I recurrently face the issue that several tracers are lost without any particular reasons (no termination value is given). One can see [it quite well in the image here](https://perso.univ-rennes1.fr/joris.heyman/img/streamtracer_bug.png), where a continuous line injection shows “holes” forming by the loss of a trajectory. It seems that the choice of the integrator between RK2/RK4 and RK4-5 modifies which tracer is missed. However, changing tracing parameters (steps, terminal speeds, max error, cell/point locator) do not have an effect.
Restarting the tracer exactly at the point it were lost do not solve the problem (no streamline is outputed).
Does someone faced the same problem ? Computing continuous trajectories is very important from a Lagrangian perspective…
Thanks a lot,
Joris
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [May 6, 2020, 3:18am UTC](https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247/2 "2020-05-06T03:18:40Z")
For reference : [https://gitlab.kitware.com/paraview/paraview/-/issues/19891](https://gitlab.kitware.com/paraview/paraview/-/issues/19891)
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [May 6, 2020, 3:18am UTC](https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247/3 "2020-05-06T03:18:59Z")
Let’s keep the discussion here.
I do see some strangeness in the behavior, but it looks like reducing the initial step length may help start the streamlines.
---
### Author:  [@Jo\_Hey](https://discourse.paraview.org/u/Jo_Hey)
#### Post date: [November 25, 2020, 7:44am UTC](https://discourse.paraview.org/t/streamtracer-loose-tracers-without-reasonfortermination/4247/4 "2020-11-25T07:44:37Z")
This effect disappeared when unchecking decomposePolyhedra while loading the openFoam case.