4 ms·
I've been running the beta for a few weeks now. The thing that stands out the most? They changed the filesystem, and I haven't noticed a thing - the upgrade tim
by jordansmithnz 10y ago
I've been running the beta for a few weeks now. The thing that stands out the most? They changed the filesystem, and I haven't noticed a thing - the upgrade time wasn't even significantly longer. If anything, the beta has been more stable than 10.2 releases.
The fact they've pulled it off so seamlessly is pretty impressive. Heck, the majority of users won't have a clue that their filesystem has changed, and that's the way it should be - users shouldn't be required to know about technical implementation.
- iancarroll 10y agoAs a counterpoint, my springboard has been sporadically crashing since I installed 10.3, and I have no idea why, but the FS change seems like an obvious culprit. Doubtful it will go away with the official release.
- r00fus 10y agoI've been on 10.3 beta and have had few crashes (fewer than since 9.x). It has been an amazingly stable beta experience for me (polar opposite from iOS5 beta which was unusable/dangerous)
- derefr 10y ago> the upgrade time wasn't even significantly longer I wonder if it works like FileVault, where there's a notion of a "partially converted" volume and blocks are slowly converted as a background-idle task after the update completes.
- mrsteveman1 10y agoCould be, it may be more like the ext3/4 -> BTRFS conversion process[1], where metadata for the new filesystem is written in the free space, but with pointers to the original data which isn't moved or re-written. [1] https://btrfs.wiki.kernel.org/index.php/Conversion_from_Ext3 https://btrfs.wiki.kernel.org/index.php/Conversion_from_Ext3
- tradersam 10y agoI believe they discussed about this being the method they were actually doing on ATP a few weeks ago, I'll find the link and report back soon.
- mrmondo 10y agoI've noticed a perceived decreased in IO latency, but haven't scientifically measured it.
- leecb 10y agoThe whole OS seems more snappy to me. Wonder if they have somehow figured out how to file access latency? Perhaps less of the filesystem metadata needs to be accessed on each file access.
- minikites 10y agoHFS+ has a global lock so only one process can access the file system at a time. I would imagine having APFS clear out that limitation speeds things up dramatically.
- charlesism 10y ago> I've been running the beta for a few weeks now. > The thing that stands out the most? They changed > the filesystem, and I haven't noticed a thing On the other hand, maybe the file system only catastrophically fails every few months.
- AceJohnny2 10y agoThat's textbook FUD. You could say that about anything.
- charlesism 10y ago> You could say that about anything. You couldn't say it about a file system that hasn't encountered a serious error in two years. The comment I replied to, sorry for my bluntness here, is pretty much "This file system hasn't give me any trouble in five or six weeks." Even if they actually verified all the files on disk, to ensure none of them are corrupt, it wouldn't signify all much, given the short duration. > That's textbook FUD. I thought it was reasonable (maybe even self-evident) but we'll probably have to agree to disagree on that.
- mandeepj 10y agoMine took way longer than earlier upgrades. Now, I am googling to see why they changed the file system. I heard about it sometime back though