Welcome to the RAS Solution Forums HEC-RAS Help HDF_ERROR trying to use HDF output file| chunk dimensions must be less than 2^32

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
    Posts
  • #18378
    alister.hood
    Participant

    I’m having errors like the one below with HEC-RAS 6.5, and 6.7 Beta, on a new machine (a desktop Win11 machine set up specifically for RAS, Core(TM) i7-14700F processor with 64GB ram). Some other projects run OK on this machine, but some give errors like this. The problematic projects work fine in 6.5 on another machine (Win11 laptop with 32GB RAM).
    Any ideas? Is there any chance getting 6.7 uninstalled would make 6.5 work?

    HDF_ERROR trying to use HDF output file
    Subroutine HDF_SET_COMPRESSION
    HDF_ERROR trying to close HDF output file:
    HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0:
    #000: ..\..\src\H5Pdcpl.c line 869 in H5Pset_chunk(): all chunk dimensions must be less than 2^32
    major: Invalid arguments to routine
    minor: Out of range

    Error starting unsteady simulation – Stopping compute process

    #18390
    alister.hood
    Participant

    I spent quite a bit of time looking at this with our IT department without success, but I think I am now making some progress on it.

    > Is there any chance getting 6.7 uninstalled would make 6.5 work?

    No, didn’t help. Also tried installing in a sandbox and various other things.

    But I ran into this problem with another project this week, and discovered that it occurred if the starting time was 01SEP2008 00:00, but not if the starting time was 00:01, or the date was 31AUG2008 or 02SEP2008.
    Eventually I found that it worked after changing the date to today, clicking compute, stopping, and changing the date back. I have no idea why changing to today fixed it, but not changing to another date… much less why it was only an issue on this particular machine.

    I also found other seemingly related problems:
    2. Before I got it “fixed”, when I ran it with a start time other than 01SEP2008 0000, the simulation started at the wrong time – when specifying a start time of 02SEP2008 0000 the simulation actually started at 01SEP2008 0240, when specifying a start time of 1100 it actually started at 0110, when specifying a start time of 1200 it started at 0120, etc. Again, this was only an issue on this particular machine.
    3. The duration of the simulation was also wrong, at least most of the time – e.g. when I specified to run from 01SEP2008 1100 to 01SEP2008 1210 it ran only ten minutes – from 01SEP2008 0110 to 01SEP2008 0120 (my guess is I didn’t look carefully enough and it was actually 11 minutes, to 0121). Again, this was only an issue on this particular machine.
    Possible 4. The simulation runs even if the specified duration is longer than my precipitation time series – I believe this is not normal – I’m sure I’ve previously had to pad the end of my time series with zeroes. Or am I mistaken and that is only for a flow boundary condition, not a precipitation boundary condition?

    Although HEC have told people that times formatted as 00:00 rather than 0000 cause errors, in all of this it made no difference whether I specified time as 00:00 or 0000.
    Also note that it made no difference saving and running a new plan.
    My rainfall series used a ten-minute time step, but I did not test to see if changing that made a difference.

Viewing 2 posts - 1 through 2 (of 2 total)
  • You must be logged in to reply to this topic.