CCDSoft2XAdaptor memory error. Error = 205.

23:28:55 I:TI1, CCDSoft2XAdaptor.CCDSoft5Camera.1: Error: memory error. Error = 205.

I had posted an occurrence of this issue with Software Bisque last week. They suggested updating to the latest SBIG drivers, which seemed to fix the problem. But TSX crashed again last night with the same error above.

This is with an STT-8300M + FWG8-STT, TSX Pro, CCDAP, and Windows 10.

I was running unguided last night, but the problem has shown up in the past with auto guiding on.

There seems like a fairly persistent problem for me, so if there is anything I can do to gather better information about the crash, let me know.

Cheers,
Joe

 

Hi Joe,
I've been monitoring your posts on the CCDAP and TSX forums and it looks like you're not finding much assistance yet. This forum seems as good a place as any to start.
I have a couple of suggestions. The first is simply a work around -- you have 6G memory, as I recall. Increase it to 8 or better, 16G. Secondly, run the Win10 Resource Monitor (Windows Administrative Tools -> Resource Monitor -> Memory tab) and watch the memory usage for the processes on your system while imaging. Watch for any obvious, large increases in memory usage by any processes over the time. Note that process name and size information over the start and end of your runs. Note any differences between when you run guided and when you run non-guided. Lastly, if you have another error of this type, zip a snip of the Resource Monitor, the CCDAP log, the TSX logs (*.txt in the Software Bisque Application Folder) and a snip of your current SBIG camera drivers (SBIGDriverChecker64 window). That should be a start. (The snipping tool can be found at Windows Accessories -> Snipping Tool).

- Rick

 

Rick,

Thanks for the ideas. I didn't know about Resource Monitor, but I will give it a try and see if I can catch what is happening.

Cheers,
Joe

 

Same error again last night:

00:03:04 I:TI1, CCDSoft2XAdaptor.CCDSoft5Camera.1: Error: memory error. Error = 205.

Firmware and drivers are all up to date.

I have upgrade the computer to 12GB RAM, which hasn't fixed the problem.


Cheers,
Joe

 

Joe,
Could you post your CCDAP log for the session in which you got this last error?
- Rick

 

Doug,

When I operate with CCDAP and TSX Pro my computer now has about 9GB of free RAM, and no signs of memory leaks as far as I can tell.

So I would agree that the problem seems to be in the code somewhere. But would that be happening in the driver, and if so, is there any way to do a stack trace or anything useful to help debug this?

Cheers,
Joe

 

Attached is the SBIG driver log from last night's failure. I don't see anything obvious, but hopefully it is useful.
The CCDAP log doesn't show anything abnormal until the error 205 gets thrown.

EDIT: The log file was over 20MB so I deleted all but the last 2MB of messages.

 

Last night I swapped out the old RAM in my computer and put in a new 8GB chip, in case the problem was in the hardware.
But I got a new and different memory error. Here are the relevant lines from the CCDAP log:

00:13:17 Lagoon Nebula, exposing...
00:23:22 I:SI1, C:\Users\Joe\Documents\Astrophotography\CCDAutoPilot\20170619_Lagoon Nebula\Data Files\Lagoon Nebula.Light.O.2x2.0E.600s.02718.fit, CCDSoft2XAdaptor.CCDSoft5Image.1: CFITSIO Error: could not allocate memory. Error = 33113.
00:23:22 Save image failed.

So at least this is a little more specific in that it seems to correspond to the CFITSIO error 113. Which program is calling CFITSIO to do the image write?

 

As Colin notes, TSX is returning the CFITSIO error to CCDAP, but mostly likely passing it through from a hardware driver given the 33133 error code. Once again this is indicative of memory starvation (as is Memory Error 205), often caused by a memory leak in a hardware driver as managed by TSX. There have been a number of threads on this topic in the TSX forums over time. Nearly all known leaks have been cleared up with the most recent software and drivers. Posting the full CCDAP log and a snapshot of your SBIG Driver Checker verifies your software and driver version levels, and helps someone understand how many images of what size and how long it is taking for your error to manifest. Watching the Resource Monitor (noted above), in particular the "CCDSoft2XAdaptor" process during the session can also betray memory leaks for this particular type of error. That evidence will be gone once the error occurs and the software resets.

- Rick

 

Rick,

Thanks for the feedback, but as noted above the lack of free memory doesn't seem to be the problem.

I have attached a copy of the CCDAP log, which will show that it captured 15 images before failing. I watched it drop the first 10 frames with, System Resources not showing any movement in memory usage. It would be helpful if System Resources produced a log file to be able to see the conditions at the time of the failure. But I had 5GB of brand new system memory free while running, to rule out hardware corruption, and it seems highly unlikely that those last 5 frames could have filled that up.

Also, I did a test a few nights ago where I captured 200 test frames in a row while watching System Resources, but there were no leaks and the run completed successfully. Unfortunately, it only seems to fail when the sky is clear.

I think I also noted above that all of the drivers are up to date, but I have attached a screen shot of that as well for verification.

Cheers,
Joe

 

SBIGDriverChecker64 tells me that I am up to date with the versions that I posted yesterday. I have attached it again with the message.
What is odd is that your settings show sbigudrv.dll as v4.87 build 2, mine shows v4.87 build 1, but build 1 has a more recent date than build 2.

 

And just to be sure, I just deleted and reinstalled the drivers, but they came back with the same versions as I posted.

 

Joe,

I’ve been studying your CCDAP log, etc. Nearly as I can tell, all your software and drivers are up to date, irrespective of the odd version numbering of the SBIG driver. You already knew this, of course.

However, my obsession with the memory leak thing stems from observations I reported to the CCDAP and SB forums last March:
http://www.bisque.com/sc/forums/p/30488/153007.aspx#153007.
I still think that it might apply to your situation. Your CCDAP log is nearly identical to ones where I generated evidence of a memory leak. I have never gotten any feedback that this particular issue was resolved. The key ingredient appeared to be the three guider images that CCDAP takes to determine proper guider exposure. The implication was that references to those images were not getting properly disposed at some point in the software stack – thus, very slow death by memory starvation.

I have not found the time to rerun those tests with the newest SBIG drivers and TSX build, as testing won’t work on simulators but must be live. But, I will soon. One thing you could try that would help is to run one of your sessions Unguided. If you still have a Memory Error, then I merely have egg all over my face and I will shut up. If not, well, maybe it gets more interesting.

- Rick

 

Rick,

I thought about the sync and autoguider images, but as I said before, I don't see any evidence of memory growing as I am watching the run. Those image are only 1.8M, so it would take over 2500 of them to fill 5G of RAM. And the 64x64 autoguider track box is far less than that.

Last night it crashed after 21 images, so it seems like a different problem.

Thanks for trying to help me crack it, though.

Cheers,
Joe

 

Yeah, does certainly sound like you have a very different issue, especially if it blows up in just a few frames. I'll keep an eye open for you on the forums, but I'm fresh out of ideas.
Best of luck,

- Rick