back to article Ever wonder why those Apple iPhone updates take so damn long?

It's a recurring pain experienced by all iPhone owners: the huge and very slow software updates that require you to plug your phone in and forget about it for 30 minutes. Many have also wondered why the phone feels the need to shut down, start up again, shut down again and start up again before finally letting you get at it. …

Silver badge

But they are 'special' guinea pigs. So it's fine to be treated that way :-)

33
3
Silver badge
WTF?

What they heard...

The company cares so much about your ongoing experience with iPhone that it had to thoroughly test the new file system so that it would be trouble-free for all our customers.

(That the method chosen would be a bowel-loosening suggestion at any other company doesn't matter, their bowels are quite reinforced, shored up with Apple's bulging egos.)

20
1
Anonymous Coward

And they got 10% back of the 30% previously lost. How generous of apple...

6
3

Except of course they weren't using users as guinea pigs. The Register with its the usual proclivity to place cynicism above information has presented half the story (or barely that).

The interesting and differentiating thing about a filing system isn't, as most assume, the physical files on the disk, but rather the metadata. So changing filing system means constructing a new file system metadata database, consistency checking it, then flipping a bit to say "now use this metadata / file indexing system." You can run all those checks including using the new filing system with a distinct process, without actually flipping the bit for the OS running on your device as a whole. It needs to be done during a system upgrade or startup process at a point where file system use by the OS using the old filing system can be prevented. So actually doing that check in a large scale release before doing it for real flipping the bit for the whole OS is eminently sensible. Of course The Register go on about wide ranging issues as though there was some specially bad upgrade experience for iOS 10.x as a whole when there was little out of the ordinary (and I get the impression they are a bit put out by that). So they are pointing to the usual low level of complaints you get at any point an OS is updated (which of course come from all over the world) as though that is proof customers were treated as guinea pigs.

But put in a more reasonable context, Apple switched out a whole f**king filing system with hardly an issue or complaint (indeed almost all users were entirely unaware) the real story is have you ever known such a large scale change go so smoothly? Go back to the 90s and early noughties and it certainly wouldn't have done. So in this nest of sneary cynicism I say "Well done Apple."

31
45
Silver badge

@SuccessCase

Apple issues regular updates = snidey Reg article about number of iPhone updates

Apple tests updates thoroughly = snidey Reg article about time taken to update

Apple doesn't explain testing methodology = snidey Reg article about playing fast & loose with data

Apple tells customers stuff = snidey Reg article about fanbois whooping cos Apple shared something

Apple doesn't tell customers stuff = snidey Reg article about Secretive Apple(TM)

Apple makes filesystem magic happen = snidey Reg article about why don't we know what's going on

Apple does business as usual but no magic = snidey Reg article about Apple lacking innovation

Basically if you like your Apple juice sour, Reg is the place to order it.

29
38
Gold badge
Devil

Re: @SuccessCase

Well, you're partly right. In that El Reg writes snidey article is basically their business model. But you're being a little special snowflake, in that they don't just do it to poor ickle Apple... In fact, can anyone think of a company they appear to actually like?*

Also they do have a point. It's all very well for these rich companies to just assume that everyone has unlimited internet, but lots of people don't. And a couple of GB could be a significant part of someone's data allowance for the month. So if Apple sorted their updates out to not have massive unneccesary extras in it, those people might be happier about it. Additionally would they be replacing any older kit that it bricked? Sure Apple are sometimes quite generous with their service, but only in Apple's usual capricious manner. As sometimes they outright break consumer laws in the coutries they operate in and reduce your consumer rights. They certainly did in the UK when I had to get 3 out of our batch of 6 iPhone 5s fixed for crap power sockets. Finally why can't they fix their fucking updates so they only download the files they need? Surely each iOS update needn't be over a GB?

--------

* Actually they're usually pretty nice about SpaceX. But then again, they're not doing product reviews of their stuff. Although I'm sure they'd be even nicer, and positively promise to kiss arse if SpaceX promised them a review model.

You can just imagine the review now. The rocket blew up, killing our reviewer. But on the plus side, nobody liked him anyway, and it made an awfully pretty explosion. Please can we have another one? 9/10 - Editor's Choice.

39
8
Silver badge

> The interesting and differentiating thing about a filing system isn't, as most assume, the physical files on the disk, but rather the metadata. So changing filing system means constructing a new file system metadata database, consistency checking it, then flipping a bit to say "now use this metadata / file indexing system."

Maybe you know what Apple has done in this instance, but in general what you say about file systems is simply not true. Obviously, you can't convert EXT3 to NTFS (or vice versa) without wiping and rewriting pretty much the whole disk, for example. OTOH if it is was straightforward a task as you say, it didn't need a multi-GB update, or half an hour of thrashing about, to do it.

9
1
Silver badge

Re: @SuccessCase

@Lord Elpuss

"Tests thoroughly" = roll out to all users?

I don't hear much support for Microsoft when they do that kind of thing.

20
2
Law

Re: @SuccessCase

"Basically if you like your Apple juice sour, Reg is the place to order it."

Good to know I'm in the right place! :)

10
0
Silver badge

Re: @SuccessCase

@ I ain't Spartacus: It's all very well for these rich companies to just assume that everyone has unlimited internet, but lots of people don't. And a couple of GB could be a significant part of someone's data allowance for the month.

It also seems to assume that ISPs and MNOs have "unlimited internet" and don't object to Apple risking clogging up their networks with test and revertive downloads to all and sundry.

7
0

Re: @SuccessCase

iOS updates will only download with a Wi-Fi connection. When using cellular data, the option to download the update is greyed out.

6
0
Bronze badge

Re: @SuccessCase

>>In fact, can anyone think of a company they appear to actually like?<<

With all the problems Scamscum have had, very rarely do Reg post a critical article, let alone an article at all. Reg always loving refers to Scamscum as "Sammy", but usually to Apple in derisive terms.

3
3
Facepalm

I say "Well done Apple."

Wow. Well, that's two minutes I won't get back. And your limited understanding of file system formatting is - well - interesting. Not to mention your apparent disregard for peoples' data put at risk by this masterful piece of hubris on Apple's part.

1
0

Re: @SuccessCase

"the option to download the update is greyed out."

.. which in software actually means that you cannot turn it off if they want to download it anyway.

Having the option available to flick, but defaulting to _disabled_ is what one should expect to see.

0
0

phone turning off at 30% is a iphone intended feature

1
1
Bronze badge

Re: I say "Well done Apple."

"apparent disregard for peoples' data put at risk by this masterful piece of hubris on Apple's part."

No, you don't understand testing. This was to protect people's data. It is quite the reverse of what you make out. Apple's testing here is minimising the risk. That's what testing does.

1
2
Anonymous Coward

Re: I say "Well done Apple."

No, you don't understand testing. This was to protect people's data. It is quite the reverse of what you make out. Apple's testing here is minimising the risk. That's what testing does.

I hope that's sarcasm.

The way to do testing is to do it with your own equipment and data first, before risking any customer data.

4
0

Re: @SuccessCase

snidey is fine, but the the register lost me years ago because of it's over-the-top anti-hard on for all things Apple. I don't see the same slant against android or other operating systems. Where's the hate for WinFS? Wasn't it part of Longhorn? Wasn't that back in 2002?

I guess you are of the opinion that Apple's switching of it's huge installed base of iOS users overnight is a minor footnote in computing history. You probably pretend Android OS deployment isn't all fragmented. I'm still waiting for my Samsung S7 to update to the latest version.

Let's read what Steven Sinofsky (ex-president of Windows division) has to say -

"I’ve lived through all the Apple migrations and all the DOS/Windows migrations and not only is this among the most feature-rich releases, it is actually running right now on my Mac (and iPhone) after an in-place upgrade. I seriously sat there watching the install process thinking “this is going to take like a day to finish and it will probably fail and roll back in the middle or something”. After about 30 minutes the whole thing was complete. The amount of amazing engineering that went into both the creation and deployment of APFS is mind-blowing. And that it was done on phones, watches, and PCs is nothing short of spectacular and except for maybe the transition from FAT to FAT32, I can’t recall anything even close to this. There are a ton of features under the covers that will surface in use of Apple devices, but mostly it will just make everything better seamlessly."

https://medium.learningbyshipping.com/wwdc-2017-some-thoughts-3ff3230cdd58

2
1
Bronze badge

Re: I say "Well done Apple."

>>No, you don't understand testing. This was to protect people's data. It is quite the reverse of what you make out. Apple's testing here is minimising the risk. That's what testing does.

I hope that's sarcasm.

The way to do testing is to do it with your own equipment and data first, before risking any customer data.<<

No it's not sarcasm at all. What makes you think Apple would not have tested the hell out of this first before trying their live test. The fact they did this live test shows how deep and careful their testing has been at all levels. Yes, there are beta testers for APFS. Testing the update live would have been the last step in a very long testing process.

That is brilliant. So to all you Anonymous Cowards out there (and I never know which particular individual I am replying to, so can't see the thought processes in your logic - I should say faulty logic), stop being blinded by your own negativity towards Apple.

Had they not done this test, you would probably be criticising them for not doing it - except you would have known nothing about it.

0
2
Silver badge

Re: I say "Well done Apple."

"And your limited understanding of file system formatting"

Formatting?

1
0
Anonymous Coward

"Trust me" = Fuck No.

I didn't agree to be Microsoft's beta tester & I'm not going to accept being one for Apple either. Since Android makers can't be arsed to fix the security issues in a timely manner, that leaves using a dumb phone instead. Damn the corporations thinking that they can use their *customers* as quality assurance. Arseholes!

44
7
Anonymous Coward

Re: "Trust me" = Fuck No.

I have regrettably needed to downvote you because you have written bad things about Apple. The swearing was pretty cool though.

12
0
Silver badge
Pirate

Re: "Trust me" = Fuck No.

"that leaves using a dumb phone instead."

I still use a dumb phone. It makes calls, sends txt msgs, and can take shitty pics. Which is exactly what I need it to do. I had a Moto X for a while in 2014 and it drove me nuts - why couldn't I get it to do what I wanted it to do and look how I wanted things to look? It's basically a miniaturized computer that also makes phone calls, and should have been cooler than any computer I'd ever had, but it was the most immutable computer I'd ever seen. I just didn't get along well with the whole "everything is locked down because users are idiots" mindset, and jailbreaking looked like a huge pain in the ass. So back to my Samsung Envy I went.

The only thing the Moto had that I really missed was maps/GPS, so I went and bought a $100 Garmin GPS, and life's been good so far.

5
0

no no no no no no no, Apple

"consistency checking it, reporting back to us whether the upgrade was 100 per cent clean and then rolling it back"

You need to be 100% sure it will be clean BEFORE testing it on my shit, Mr Apple. And if you were 100% you would have no need to test it on me would you?

This is NOT the Apple way.

21
8

Re: no no no no no no no, Apple

Au contraire. It clearly is the Apple way.

45
4
Silver badge

Re: no no no no no no no, Apple

I think you're unclear on what they did. They upgraded, checked and rolled back. The check was to see if the migration succeeded, but the upgrade didn't mess with the existing filesystem it at all - it created new metadata, and left the existing metadata and file data in place.

There was no risk with what they did, other than it taking a lot time and running down your battery if you upgraded while on battery (and presumably had a bad battery, since it won't upgrade on battery if you have <50% battery remaining)

20
19
Bronze badge

Re: no no no no no no no, Apple

"There was no risk with what they did" - are you sure? What's the point of testing if the outcome is certain? Now, the remaining questions are if all upgrades involved the test, why it was rolled back despite being successful (to wear Flash memory?) and what was the risk of failure (of the whole process including rollback or the rollback alone) and who would be liable for data loss. That said none of my upgrades took this long, possibly because none of this testing could be performed due to the limited free storage or maybe because I'd disabled the telemetry upload (and hopefully Apple as opposed to MS respects user's setting).

26
6

Re: no no no no no no no, Apple

@DougS - "There was no risk with what they did, ..."

"And it likely explains why Apple had to put out a 10.1.1 update just a week after the massive 10.1 update, *because people were continuing to report serious problems*."

28
4
Silver badge

Re: no no no no no no no, Apple

The point of testing was to do the consistency test, to see if there are any filesystems 'in the wild' that have factors the conversion software failed to account for. There are probably a few phones in the wild that have clocked 5+ years of consecutive use. The Apple employees get new ones every year, so they can't count on seeing everything that the other billion iPhones have.

As for why it was rolled back if successful, they might not have been ready to enable the filesystem permanently. I would guess the conversion WAS permanent for some Apple employees, so they could test the new filesystem before they rolled it out everywhere.

The 10.1.1 update thing was pure speculation on the part of El Reg. It isn't rare to see a release have problems, like 10.1 did with losing health data for some people. That doesn't equate to "oh noes, evil Apple experimented on us and it went wrong, that's why 10.1.1 was released!" Not sure why everyone is claiming this upgrade testing is an issue. When have a billion filesystems ever been upgraded without any issues before? That's a pretty impressive achievement, so I wouldn't be so quick to question the methods Apple used. They worked.

14
19
Silver badge
Happy

When have a billion filesystems ever been upgraded without permission before?

FTFY

17
4
Silver badge

Re: When have a billion filesystems ever been upgraded without permission before?

Hey, no problem. I want to "test" Ext5. Can I just use your phone as a test device before I roll it out and not tell you?

Tinkering with filesystem code is the easiest way to cause data corruption. It's not like upgrading to a new style of GUI widget, it's literally implanting offsets and pointers into a permanent data store, that contain the locations of your files in storage. Get that wrong - even one bit wrong in an extreme circumstance, and you can trash the root or even end up with a slowly-corrupting filesystem.

Even with all the checksum checks and error correction in the world (hint: What are you going to do if the checksum comes back as "wrong"? Invalidate data. What if that checksum wasn't ACTUALLY wrong, but you forgot to include new-fancy-attribute-X in the data you were checksumming. You just lost a sector. Error "correction" does what it says - fixes things that it believes are errors.).

Using people as filesystem guinea pigs is the worst thing you can do in terms of data integrity. Even if just for testing. One wrong pointer and you can wipe out even the "backup" or other partitions - because they are literally just numbers in a box and if you have two "partitions" you've just assigned a convention of using numbers between A and B as one partition and between B and C as another, and then recorded A, B and C somehow. Get that wrong - even one bit out, or fail to check the write properly or any number of events and you trash the "new" trial partition and the "safe" backup of the user's data in one hit.

19
6

Re: no no no no no no no, Apple

No risk? this update bricked my girlfriends iPhone quite spectacularly, took forever to get the thing to boot again after re-flashing through iTunes.

16
3
Silver badge

Re: no no no no no no no, Apple

"There was no risk with what they did"

If there was no risk, then it didn't need testing. Being pretty confident you can make a change and then roll it back with no problems does not justify doing your testing on a live system - the whole point of testing is to change that "pretty confident" into "absolutely certain", at which point your testing is done and you can roll things out properly.

And of course running down batteries wasn't the only affect on users even where things went fine. Many people still have download limits on internet connections, on land lines as well as mobile. Forcing your customers to download a GB or two for no good reason can cause real inconvenience - a family with a few iDevices could easily find themselves using up half their monthly allowance, as well as being unable to make or receive calls and such for an hour or so. For most it will be just a mild inconvenience that they shouldn't have been forced to suffer, for some it could be a lot more than that, and in neither case is there any excuse for Apple to have caused it for no reason and with no warning.

14
2
Anonymous Coward

Re: When have a billion filesystems ever been upgraded without permission before?

I felt a great disturbance, as if a billion filesystems were upgraded, and were suddenly silenced. I fear something terrible has happened.

15
0
Silver badge

Re: no no no no no no no, Apple

Now, the remaining questions are if all upgrades involved the test, why it was rolled back despite being successful (to wear Flash memory?)

I suspect:

- 10.1 only supported HFS+, but contained the filesystem migration utility and test program.

- 10.2 also only supported HFS+, but contained the final filesystem migration utility (and test program?).

- 10.3 only supported the new APFS, as part of the update it invoked the preinstalled migration utility.

and what was the risk of failure (of the whole process including rollback or the rollback alone)

Properly thought through, ie. allow for failure and thus rollback at any point in the process, the risk is low.

and who would be liable for data loss.

You do have sync and backup to iCloud enabled?

3
1
Silver badge
Childcatcher

Re: no no no no no no no, Apple

The Apple employees get new ones every year, so they can't count on seeing everything that the other billion iPhones have.

So Apple doesn't keep old hardware on hand for testing? They don't have an automated test process that would allow multiple test scenarios to play out? They don't have any sort of virtualized test environment?

Two issues stand out to me: lack of transparency concerning what they do and lack of informed consent from their customers to do it. There are a lot of good things to be said concerning Apple's products, but their practices leave a bit to be desired.

12
2
Silver badge

Re: no no no no no no no, Apple

Again, the 10.1 and 10.2 'upgrades' didn't touch any data or metadata in your filesystem. It is as if I upgraded your NTFS filesystem to ext4 by using free space to create inodes that pointed to the existing data blocks from your NTFS filesystem. Your NTFS filesystem would be intact, but I could do some various tests to see how well the ext4 upgrade went. Once I'm done I wouldn't have anything to do - the blocks containing those ext4 inodes would still be listed as free space in your NTFS filesystem. There's no "rollback" required for this because it doesn't do anything that needs to be reversed.

Now obviously when 10.3 came around it was the equivalent of starting to use the ext4 filesystem, and dumping the NTFS metadata (making it free space) But by this point they had fully tested the migration so they knew it worked, and the filesystem itself had been used on Macs for a while so they knew it worked. Hence the complete lack of disaster from the 10.3 upgrade/migration.

As for keeping old hardware on hand, there's a difference from having an iPhone 5 laying around you can experiment on, which they obviously have, and having an iPhone 5 that has been in daily use since launch. As anyone who has every deployed anything knows, end users are really good at having scenarios that developer/administrator testing doesn't find. Maybe the upgrade testing was unnecessary, as no problems were found. But better to find it out during such testing, which is has no possible ill effect because the actual filesystem isn't being altered, than to find out there's a corner case or two they didn't account for or know to test for.

4
5
Silver badge
FAIL

Re: no no no no no no no, Apple

"I wouldn't be so quick to question the methods Apple used. They worked."

THIS TIME. Lucky once does not validate a stupid decision. Ever hear of Russian Roulette? Five out of six times that goes without a problem as well.

- If they felt the need to test it, then there was a chance it would bomb.

- They did not notify or ask the userbase about this.

Apple thinks it still owns the devices you carry and has full rights to manage them anyway they want. That's bullshit. FAIL.

11
3
Anonymous Coward

Re: no no no no no no no, Apple

"Again, the 10.1 and 10.2 'upgrades' didn't touch any data or metadata in your filesystem"

Sorry but that didn't, should be shouldn't. The upgrade test would have needed to directly access the storage, bypassing the current file system, if it didn't bypass the file system, direct access would have needed to be implemented in the current file system, therefore testing this also. If that didn't take into account of these possible differences you mentioned for the reason of the metadata test then it could have caused corruption while writing directly to this possibly free space.

If they just wrote directly to the flash then they would have needed to directly update the allocation table, if there was a problem in this code, data corruption could occur.

If they just wrote meta data to a file on the current file system, checked that content, thus using the current file system, then no corruption should occur. But then they wouldnt have been testing the actual upgrade process or testing those fringe file systems you gave as to the reason for the test.

1
2
Silver badge

Re: no no no no no no no, Apple @DougS

I don't know whether you're not thinking this through, don't really understand the differences between different filesystem types, or are just naieve.

It is not easy to, say, do an in place conversion from EXT3 to NTFS. Everything from the tracking of free space, block and fragment allocation and metadata are different between the file systems, meaning that to convert the filesystem it will require every file to be read and re-written. This will effectively destroy the original filesystem while creating the new, meaning that a roll-back is as intensive and risky as the conversion.

Now if the changes between the filesystem types are evolutionary rather than revolutionary, it may be possible to do an in-place upgrade. So, it is possible to upgrade from EXT2 to EXT3, because most of the filesystem structures are the same or very similar. The same is true of EXT3 to EXT4. But these are a family of fileystems, designed for backward compatibility.

If APFS (I'm soooo glad they did not call it AFS, which has been used at least once already) keeps the files in place, and just creates new metadata in free space, as you possibly suggest, it would almost certainly be possible to do this without touching the original data or metadata. But does something like this actually count as a 'new' file system, rather than a new version of the old filesystem?

I would also be interested in how much wear the flash memory will suffer from repeated writing during these test upgrades.

4
1

explains why they bricked my Iphone 4

A while back an upgrade bricked my Iphone 4, and there's no way to roll it back (and I hadn't saved the required blobs first)

Nice one Apple. Fuckers.

3
1

Re: no no no no no no no, Apple

"There was no risk with what they did" - are you sure?

--Yes, based on the description of what they did.

What's the point of testing if the outcome is certain?

--The outcome of the test was not certain. The outcome to the end user (no change to the in-use filesystem metadata) was certain since they would not apply the changes.

why it was rolled back despite being successful

-- Because again you misunderstand what was being tested. They didn't want the active filesystem to change in that update. They were checking the updater itself.

(to wear Flash memory?)

-- get real. The flash in those devices aren't going to wear out.

and what was the risk of failure (of the whole process including rollback or the rollback alone)

--if they never actually changed the active filesystem and only practiced building a new filesystem metadata there is little risk as long as they checked for enough space, didn't overwrite real data, or made sure the changes were separate from real active data.

and who would be liable for data loss.

--you should have backed up before the upgrade anyway. You backed up the iPhone first, right?

2
6
Anonymous Coward

Re: no no no no no no no, Apple

this update bricked my girlfriends iPhone...

I'm told Apple has now adjusted the fs migration code to cater for a phone full of duck-face selfies. Her phone should be ok from now on.

1
1
Bronze badge

Re: When have a billion filesystems ever been upgraded without permission before?

>>Using people as filesystem guinea pigs is the worst thing you can do in terms of data integrity<<

At some stage they need to update to the new file system. The worst thing they can do is just install it and then find a whole lot of the user base has something strange that stops it working. That is the worst thing they can do is not test it.

This is probably the best thing they can do and must be the biggest test in the history of the industry.

Instead of the Register's spin and criticism here, Apple should be commended for doing such an extensive test.

1
2
Silver badge

Re: no no no no no no no, Apple

>who would be liable for data loss.

Well you would, obviously.

Wait, you didn't back up your data before doing an OS upgrade? No silver el-reg badge for you!

During an OS upgrade, when you might conceivably get a file system upgrade anyway would be the perfect time to do some testing. Would you prefer it to be rolled out without a roll-back plan?

If I were rolling something out, I'd want to do some real-world unit testing too.

I rather despise most of Apple's business practices but this isn't one of them.

0
2
Silver badge
Paris Hilton

Re: no no no no no no no, Apple

>No risk? this update bricked my girlfriends iPhone quite spectacularly, took forever to get the thing to boot again after re-flashing through iTunes.

So... not bricked, but requiring a restore from backup. It was an OS update...It was done with permission and suggesting that it was the fs test which borked it seems like a bit of a leap of logic.

Does anyone here work in IT?

0
4
Anonymous Coward

Re: When have a billion filesystems ever been upgraded without permission before?

... these are not the 'droids you are looking for.

0
0
Bronze badge

Re: no no no no no no no, Apple

"I rather despise most of Apple's business practices but this isn't one of them."

What do you despise and why?

1
1

Re: no no no no no no no, Apple

I'm sure Apple sent out a pre-alpha version of the upgrade to users. /s

0
0

Re: When have a billion filesystems ever been upgraded without permission before?

Man you're all a bunch of bitching cry babies. Unless you work for Apple, you can't tell me they didn't test this out hundreds of thousands of times in a lab. You also don't know if they tested it on users with a full iCloud backup of their phone.

All you know is to attack a firm that updates hundreds of millions of phones en masse, annually, and yet Android is more fractured than Evil Knievel but get's a pass.

2
1

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Forums

Biting the hand that feeds IT © 1998–2017