Did New Build for Mac Break WBPP?

jtsantes said:
The last Time Machine backup was last night at 6:20pm. Having said that, for some reason, when I am doing a WBPP or Fast Batch processing operation, I get "temporary files" in my Library folder "Clouddocs>Session>i folder. These files take up GB worth of data and I always delete them after a wbpp or Fast batch process. I am wondering whether those files might be screwing everything up. But again, Batch processing only started failing since the latest Build 1614 was downloaded a couple weeks ago. This build was specific for Mac users and was supposed to fix some problems. I noted in Fast Batch that it selects 5 reference frames, not all. Is there a way to select fewer frames in wbpp by changing some default or setting?

Is there anywhere or anybody else that might have a solution. I have an astro buddy that has had this same sort of failure "occasionally" but not enough to cause him any concern. He runs a Mac, but one using the M2 silicon. My laptop uses the M1 chip. Please let me know what you think. The real issue is time. If I try to drizzle in wbpp, I might wait 4 hours and see a failure, so it is annoying.

Thanks for your help.
Click to expand...
There are also file I/O errors (line 47866) because your disk is full in the middle of processing.
 
yeah you are running with icloud desktop and documents turned on. so there is a cloud sync process running. just move all your files to some other place in your home dirt and don’t target a folder in your documents or desktop folder.
 
In fact, when running PixInsight, don't have any files (image data; database data; scripts; ...) in a folder that has a backup / mirror running on it (either to cloud or to local backup device)...
... and don't say "but it always worked before"; it may sometimes work now - but you can be sure that the time it fails will be 14 hours into you 1000-subs session.
 
cloudbait said:
The fact that your temp files aren't getting cleared also suggests some kind of background routine that is moving files around. And anything in Clouddocs... well, there you go. I think that's almost certainly your problem.
Click to expand...
I normally keep all the files on my desktop prior to processing and then move everything associated with that image to my external SS driveI when I'm done processing the image. I will try to load all new files directly into my external drive and then select the master light to perform a wbpp rather than keeping the master light on my desktop. I'm not sure that will help but it's worth a try. I can try doing a wbpp on a small set of subs to see whether the temp files are in the Clouddocs folder.

Having said that, do you know of a setting in wbpp to select how many reference frames that operation selects? As I mentioned, the Fast Batch operation selects 5 frames.
 
jtsantes said:
I normally keep all the files on my desktop prior to processing and then move everything associated with that image to my external SS driveI when I'm done processing the image. I will try to load all new files directly into my external drive and then select the master light to perform a wbpp rather than keeping the master light on my desktop. I'm not sure that will help but it's worth a try. I can try doing a wbpp on a small set of subs to see whether the temp files are in the Clouddocs folder.

Having said that, do you know of a setting in wbpp to select how many reference frames that operation selects? As I mentioned, the Fast Batch operation selects 5 frames.
Click to expand...
I'm not sure what you mean. WBPP selects one reference frame. Nor do I quite understand what you mean by selecting the master light to perform WBPP.

All you need to do is make sure that any folders PI writes files to are not being synced to iCloud, and not being automatically backed up in real time. This includes the folder PI uses for creating temporary files. This will prevent the creation of the corrupt files that are crashing your WBPP run, and it will prevent the massive count of temporary files that is filling up your disk and creating additional errors. You can do that by choosing different folders, or by reconfiguring your sync/backup utilities.
 
I think he is probably referring to to the number of frames used to generate the LNorm reference in which case its in this section of WBPP. Drop the default of 20 to 5.
 

Attachments

  • Screenshot 2024-11-08 at 15.08.17.png
    Screenshot 2024-11-08 at 15.08.17.png
    562.9 KB · Views: 14
chris.bailey said:
I think he is probably referring to to the number of frames used to generate the LNorm reference in which case its in this section of WBPP. Drop the default of 20 to 5.
Click to expand...
Does FBPP even do local normalization? (I've never even opened the script, so I know nothing about its details.)
 
Since posting my original question regarding WBPP failing the first time I run it, I think I have solved the issue. I have placed all of my subs in my external SS drive and created the Stack in WBPP using a folder also in that external drive. I stacked 2 sets of different subs and WBPP did not fail. In addition, interim files were no longer stored in the Clouddocs folder in my library, so there was nothing to delete. Thank you Cloudbait for bringing that "sync" issue up and an earlier thread.
 
cloudbait said:
So what reference frames are you talking about? WBPP only picks one, because you can't have more than one. Surely FBPP does the same?
Click to expand...
In WBPP, one of the operations is "Reference Frame Selection." I believe that the LN operation uses that and if it is not a good frame, LN fails, which is what has been happening. FBPP does not do a LN to my knowledge so it cannot fail. Perhaps I am wrong, but it is the common link that I have found. Believe it or not, I can tell whether wbpp will fail judging on how long it takes for the reference frame to be selected...10 vs 40 (+/-) ms.
 
To all who provided input for this topic: Thank You. I'm still a newbie to AP and especially PI, so I appreciate all the help I can get from long-time PI users and experts. At 78 years old, I'm guessing I'm older than most of you, but my goal is to do as much as I can in AP until I join my ancestors in the stars. Hopefully, my sometime-newbie questions don't annoy you. Thank again, and clear skies to all.
 
jtsantes said:
In WBPP, one of the operations is "Reference Frame Selection." I believe that the LN operation uses that and if it is not a good frame, LN fails, which is what has been happening. FBPP does not do a LN to my knowledge so it cannot fail. Perhaps I am wrong, but it is the common link that I have found. Believe it or not, I can tell whether wbpp will fail judging on how long it takes for the reference frame to be selected...10 vs 40 (+/-) ms.
Click to expand...
I see. Well, the reason it was failing (as can be seen in the log file) is that the file was corrupt, presumably because of the cloud syncing, which you've now addressed.
 
You must log in or register to reply here.
Back
Top