elecstorm

Houdini installed per node and on the workstations. We almost always use bgeo.sc. It's debian 8 Jessie, on blades, workstations and on the old storage. The newest storage is running scientific linux with custom build. We will check the descriptors limit on the storage, but the descriptors are not limited in my system.

LaidlawFX, thanks for the answer. I really think it's software problem. Different storages, different blades. Nuke creates the same load on the storage, but we don't see any data loss. I was hoping that there is a low-level setting in Houdini for writing geometry files.

Hi! I have a strange issue. When I write many files at the same time, some of them breaks. It can be 200-300 files simultaneously size of about 10 Mb, or 70-80 files of 100 Mb - I have the same result: files exists, but 5-10% raise read error in Houdini. They have a normal size or less than normal on the storage. I tried this on the old storage and on the newest storage (with perfect access time, speed, caching etc) - the same result. Houdini 16 behaves just like Houdini 15.5. When I starting write this files, we have write errors on the storage from other software like Nuke or just write files from the OS. Another picture is not observed: when we write hundreds exrs from nuke - there is no file breaks on the storage. This occurs for many months, and I think the reason is in the method of writing files from Houdini. Is there a technology to change the method of writing geometry from Houdini (flags, variables)?

Hi! I have the same question. How can I get HDA position or position where it will be created in OnCreated Event? kwargs['node'].position() returns [0, 0] as expected.

Hi! Is there a way to bake a volume to point cloud during render? I did it for polygons by pcwrite function and by micropoly mantra with uniform measuring (1 point for 1 primitive). But if I do this with any volume shader - pcwrite does not work after pbrlighting node (as part of compute lighting node, and I did pbrlighting manually to get Cf): rendering freezes. I want to get point cloud with Cf color for each voxel.

That's unreal. Naming in, naming out, naming in, naming out...

Ok, thank you! I checked script with hython - it works.

Ok, I found this way: try: hou.node("/out/mantra1").render((1,1,1)) except hou.Error, ex: print ex.instanceMessage() It works in the python shell, I have not tested it with the hython. But I don't understand this exception (I found it somewhere): except hou.Error, ex "hou.Error, ex" - It's not definition (because of comma). So, what is it?

Hi! Is there a way to catch Mantra error in python script? I want to implement something like this: hou.hipFile.load(hipPath) try: hou.node("/out/mantra").render((1, 1, 1)) except error: #launch manta again or something else... hou.hipFile.clear() print "Frame " + str(1) + " was rendered"

It is strange, because source volume in SOP level also has repeate border type (and volumegradientfile sample it correctly). Hip: https://www.dropbox.com/s/22z2wccgjbbfhx9/gradient.hipnc It's hard - with locked volume. Now I use two fields: one with repeate border for sample this field out of it's boundary. And the same volume with streak border to calculate the gradient field. But I want to understand why it's happeinig.

This artifact appears with border type "repeat" of the source volume (I use sop scalar field to get it from sop). With mode "streak" it's ok, but I need to infinite repeat this volume to use it in the simulation(.

Hi! I want to calculate gradient field in dopnetwork. It works perfect, when I do it in gasFieldVop using volumegradientfile node. But when I do the same gradient using gasAnalysis in gradient mode, I have some artefacts on the border. How can I solve this problem?

Ok, deleting pieces is not problem, but time offset will not work, because each piece will have to be frozen separately. Here is an example. Fracture object's peaces are deleted during the simulation, but they are deleted from the scene also... delete_example.hipnc

It's fine solution, thank you. I can use delete-node instead of swtich. But freezing object by timeshift works with individual objects only. What about fractured object with hundreds of pieces?

Hi All! I have a question. There are many objects in scene. I need to simulate some groups of them in few frames (about 1-2 seconds) - using DOP like a conveyor of sumulation. I need to add objects to DOP and delete them during the simulation. Node "active value" works fine, but incative objects anyway slow down simulation's speed. Is there any solution to remove objects from DOP during simulation? And freeze object's last state in scene.