Every news item that comes out is like a feeding frenzy for the piranha
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
APFS Bug in macOS High Sierra Can Cause Data Loss When Writing to Disk Images
- Thread starter MacRumors
- Start date
- Sort by reaction score
Apple doesn't really fix most bugs because they just release a "new" version and the sheeple move on...Lion comes to mind for me.
right...thats why server grade expensive memories have hardware features like ECC, to protect performance & data loss...
Well, yes and no. Yes in that several of the responses in here are unwarranted — the bug is, for example, irrelevant to the automatic conversion that took place.
But no in that, ironically, the bug is particularly evil if you use a backup tool, because that tool is particularly likely to use sparse disk images. It's even more evil in that even tools like md5 can make you believe everything went fine (presumably due to in-memory caching).
Apple did a good job converting hundreds of millions of devices to APFS. Apple also did a poor job making various macOS features work well in APFS. In part, they acknowledge that — such as by not permitting APFS on Fusion Drive — and in part, they don't seem quite aware of what works well and what doesn't.
Yup.
[doublepost=1519045291][/doublepost]
My guess is you're using the term "memory leak" incorrectly.
(In fact, generally speaking, memory leaks have absolutely nothing to do with data loss.)
We're not talking about memory leaks, and ram, but rather a bug in the file system that is causing data loss on write operations
I too would be surprised. First, APFS formats only on drives that are SSD. It won't format on Fusion drives, nor will it format on standard platters. So the only people who will encounter this bug are those who have an all-SSD drive who try to create a sparse disk image on it. Yes, it's a serious bug (which Apple has patched for the next update), but it would be evident only in a limited number of special case uses.
Seems like more stuff I am happy I don't have to deal with on iOS. Hoping for a revitalization of macOS at some point, but I have moved on for now.
i know... but someone was so is it against the rules to talk about different things..
"I feel your pain. Had a major memory leak after moving to Sierra, let alone High Sierra. Progressively unresponsive about 10 minutes after reboot. Wasn’t the notorious iTunes leak either. Couldn’t trace it. Went back to El Cap and am quite happy. My only complaint is that Photos can’t pick a new key frame from a live photo. I’ll live."
You realize that ECC memory actually reduces performance? This feature is to prevent bitrot at the COST of performance.
No consumer product uses them. It's like expecting your Tesla to come with military grade tires just because it's expensive.
Last edited:
Let me just ask you - how do you recognize memory leak? I think you don't really understand what you're talking about.
I do agree, but then it comes back that it could have been in internal testing for a decade (and maybe it has, we don't know how long they worked on it) and perhaps no one would have ever come across this set of scenarios.
Mike, who works extensively with sparse disk images (as do I by the way and talk to Mike quite a lot) has had access to APFS since the first Sierra beta, so that's well over 18 months and he's only found it this week.
Also unfortunately it's the case but I doubt Mike would have tested it it properly in beta and neither would I, I haven't found the APFS issues I have until I decided to use it for all client systems.
wow... read it again "features like ECC" which means there are other hardware features in those expensive memories similar to the ECC..so as you see i didnt say ECC will improve performance or prevent performance loss...
*and the key thing is that the drive would have to be full as well. Not many people are using their system drive full whilst also trying to backup to a full system drive.
You'd basically need a HFS formatted network drive that you didn't pay any attention to the capacity off, using an APFS sparesdisk image to backup to on with CCC itself (or perhaps your own rsync script) and that disk would need to get full and then you'd reach the bug, but you wouldn't actually realise it until you'd come to get something from the sparsedisk image which was corrupt.
Also you'd expect in this kind of setup the person would at least have some kind of notification for when their NAS drive was full.
Mike Bombich is a nice guy although I’ve never met the man personally. I use CCC all the time for all my backup and disk image purposes and the software is rock solid. I don’t know why someone would make fun of his name.
This is a serious bug and I’m surprised Apple didn’t test for it because it is really a core of APFS’s new feature, space sharing. That said, in everyday life hopefully people aren’t filling up drives. The only drives I come close to filling are backup drives. And I’m not writing to a sparse bundle on them. Nor are they APFS.
We knew APFS had some problems since they couldn’t get it to work properly yet with Fusion drives. So we knew this was an incomplete file system.
I guess one good thing about HFS+ was it was so old all the bugs were worked out before half the customers were born.
We knew APFS had some problems since they couldn’t get it to work properly yet with Fusion drives. So we knew this was an incomplete file system.
I guess one good thing about HFS+ was it was so old all the bugs were worked out before half the customers were born.
I don't want to be "that guy" but I assume English isn't your first language, so just a polite tip that "memories" isn't really required. Memory, works as the plural too in this situation.
"which means there are other hardware features in expensive memory"
Would you please list at least one such feature?
I didn't want to bring this up, but I know quite a bit about computer hardware, server grade to. I can tell you that there's not even one such feature. ECC is basically the only differentiator.
Well said.
Given that the original report defined the problem very well, the sheer amount of disinformation being posted here is staggering.
Very few average users will encounter this problem simply because they don't use 'sparse disks'.
Also, very few commentators here have ever had experience of developing a filesystem. Even simple ones are quite complex. I wrote one in 1975 and believe me it isn't easy and that's before you get into all the locking needed when you get to the multi-threaded, multi-tasking operating systems we have these days.
TBH, I think that Apple has done a great job with AFPS Devices that use iOS won't see this bug and I recon that 99% of Mac users won't see this even if they choose to use Sparse Volumes.
Can we reduce the amount of Hot Air being expended?
They have?
I wouldn't even have a risk of this bug impacting me if the conversion to APFS was not automatic. I only have High Sierra on an Air at home. I was happy with HFS+.