Full disk causes file deletion and corruption
| Originator: | oneapplesomeapple | ||
| Number: | rdar://19112893 | Date Originated: | 16/11/2014 |
| Status: | Open | Resolved: | |
| Product: | OSX | Product Version: | 10.9.5+ |
| Classification: | Major | Reproducible: | Consistently |
There is a major bug in Apple’s OSX Mavericks Operating System. Whenever the disk is approaching full, the OS will delete (or loose track) of files which are open in running applications. If the application crashes because of the low disk space, the file is lost permanently to the Operating System (until a disk carving method such as PhotoRec can be used to ‘undelete’ the file). Some applications will handle the situation gracefully, refusing to update an existing file, or refusing to create a new file where requested. The low disk space creates stability issues for some memory intensive applications, notably Pages and Xcode, which often crash when the situation occurs. However, for whatever reason, files in use by applications do not seem to remain on the physical storage. For example, a file in use by (being edited within) Pages cannot be located within the Finder- it does not appear to exist, and for obvious reasons, fails to save with a message simply stating “This file cannot be saved”. In another example a file open in Xcode, which crashes entirely in this situation, will not be found within the finder, nor within the Project Navigator when the application is relaunched, even if the low space issue is resolved. The file effectively appears to be deleted by the computer. This issue is consistently repeatable, occurring every time the remaining disk space becomes low. It has been observed to affect OSX Mavericks 10.9.5+. Neither the application, nor the Operating System, correctly handle the situation. It would be expected that a file previously saved to the physical storage remains there, regardless of any low space condition, and any request to update or overwrite the file be simply refused if cannot be completed. Certainly it is not expected that the file be removed from the physical storage simply because there is no more space to update it.
Comments
Please note: Reports posted here will not necessarily be seen by Apple. All problems should be submitted at bugreport.apple.com before they are posted here. Please only post information for Radars that you have filed yourself, and please do not include Apple confidential information in your posts. Thank you!
There was a first draft of this report which called some Apple Devs a bunch of sweaty, overpaid, short sighted cockspankers, but it has since been rewritten. The answer to making mistakes is not to simply try and build something larger.
It is not enough to claim that a backup system should have been in place (indeed one was), nor is it enough to claim that use of the software comes with no liability, nor is it enough to claim that the latest software updates should be applied- the recent OSX Yosemite is less than a month old (for General Stable Release), and you all know that it is important that development, production and responsible machines run the most demonstrably stable version of the OS which accepts the versions of the software required for use. In fact it is generally accepted that no effort will be spent in even testing an OS upgrade in a virtual test copy of a responsible machine until the target version has reached at least a month of stability in the whatever.1 (or .0.1) version.
Discovery and recovery of this bug occurred within the last few days of collating the necessary paperwork for filing patents, requiring significant recovery effort and risking the loss of being first to file.
Recently, a Class Action law suit has been brought in the United States against Red Bull, noting that their claim of a superior product (and associated premium pricing) was unjustified due to their lack of exceptional recipe.
Apple claims their work to have resulted “The World’s Most Advanced Operating System”, and “The World’s Most Powerful Operating System”.
I know there are many people who would agree that this is in fact not the case- rather due to (sometimes artificial) restrictions on compatible version numbers and hardware Apple do not empower but restrict.
Similarly I know there are many people who, despite using Apple products, do not trust Apple to safely hold their data, nor that they may be able to access their work or data in the near future (or equally that all their hard work may even be useable once the development has finished) thanks to these restrictive updates. Conversely I’m aware that nothing should ever be expected and that life carries an inherent risk, (however wouldn’t it be disappointing to learn that such a fantastic platform we’ve created for ourselves - and for no other reason - may simply be valueless ahead of it’s time).
Therefore it seems that a similar case could be considered against Apple, and it might go something like this: Apple, through use of slogans such as those above claim a clear superiority of their software and hardware products. Yet through a demonstrable lack of correct basic implementation of features & capabilities and deliberately (and unnecessarily) restrictive policies despite an aggressive update strategy they do provide the empowerment or technological advancement on which they claim their premium is based.
If there is anyone interested in making this case in the “US”, in “Europe” (Group Action) or “Africa, Middle East, and India” (anywhere we would be able to pursue such action), please contact the email address of this post (oneapplesomeapple -at- outlook -dot- com).