Reloaded boundary files do not update in time

[mcgratta]

PMMA_60_kW_4_cm.txt

Run this 12 mesh case and while it is still running, open Smokeview. Load a boundary file. Note the time of the last frame. Unload it, wait a bit, then reload. In my experience, the time does not advance. It does for slice and other types of files.

[gforney]

Ill take a look

[gforney]

where is fds running and where is smokeview running

[gforney]

ie on spark or your PC

[mcgratta]

Job runs on spark and SV on PC via samba

…

On Thu, Sep 5, 2024, 8:50 PM gforney ***@***.***> wrote: ie on spark or your PC — Reply to this email directly, view it on GitHub <#2009 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACWPCFYOZTHFNHAM7WLSCKDZVD33TAVCNFSM6AAAAABNXFCPXWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZSHE3TGMRZGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[gforney]

I see what you are seeing when I run on spark and view on my pc using samba. at home I was not seeing the problem when I ran on my pc and viewed on my PC. I'm trying to run on my pc at work and am getting errors in attached file . I ran with debug version of FDS-6.9.1-825-g48a6b66-master . think the important lines are:

system msg for write_line failure : No error
forrtl: severe (157): Program Exception - access violation
Image PC Routine Line Source
impi.dll 00007FFE1CD636CE Unknown Unknown Unknown
impi.dll 00007FFE1CDFF437 Unknown Unknown Unknown
impi.dll 00007FFE1E1561DD MPI_INIT_THREAD_F 21 init_thread_f08ts.f90
fds_impi_intel_wi 00007FF67778E345 MAIN__ 89 main.f90

errors.txt

[mcgratta]

My fault. I propbably gave you a test file. No matter. I think you see what the problem is. This only seems to be a problem with boundary files.

[gforney]

this looks to be a problem with samba. i don't see the problem when I run smokeview on windows looking at a case running on windows or running smokeview on spark running a case on spark. I only see the problem when I run smokeview on windows viewing a case running on spark (same case that worked with the linux smokeview). I'm looking to see how smokeview is handling boundary files differently than slice files

[mcgratta]

Is there some sort of timer file that is not being updated? Does sv still write out some kind of temp file indicating time of last read? Does sv write any kind of file, because if it does, it may not have permission.

[gforney]

smokeview determines the number of time steps for slice files by "walking" through the file by skipping over the header and each time frame (which is of constant size) until it gets to the end of file, counting the number of data frames along the way. this works when files are in a samba share

smokeview determines the number of time steps for boundary files by determining the size of the boundary file, subtracting the header size then dividing by the data frame size (like slice files each boundary file data frame is the same size). the problem is that the call to determine the size of the file returns the wrong value

as we discussed I'll leave this alone (ie not try to change samba) except I'll see if I can find other ways to find the file size which may work