OS X 10.9.3 Is Toxic | Hacker News Reader

OS X 10.9.3 Is Toxic

brendangregg.com

123 pointsakerl_135 comments

In an attempt to balance the discussion, I'm on 10.9.3 and I've had absolutely zero issues so far.

Same, I run 10.9.3 on five machines - 6 year old Mac Pro, a 2 year old 13" MacBook Pro, a 2 year old retina MacBook Pro and a 6 month old MacBook Air.

I use AirPlay, external monitors and projectors all the time. No problems here.

As we have to remind ourselves at work, hundreds of thousands of happy customers don't complain, but the hundred who do (quite rightly) are very loud.

That's my experience on my two Macs.

In fact, to the contrary, it seems to have finally fixed my issue where Mail would randomly stop fetching new messages from Gmail until restarted. At least I haven't seen the symptoms of that in a while.

It going to take me months/years to trust Mail again. Such a horrible bug for them to release with.

10.9.3 has been pretty rock solid for me on my Mac Pro, as has every OS I've had on it since Leopard (10.5), although I skipped right past 10.7 to 10.8.2.

My Air has had some weird won't-wake-up issues from time to time since 10.9.3 came out. Maybe its a laptop thing?

I honestly haven't seen a kernel panic in years though on any of my computers though. The last time I did it was because I had a slew of bad sectors on my HDD.

I was just thinking I should write a daily blog for the three different Macs I interact with daily and the problems they don't have. I even do remote updates to a collocated Mac Mini that is just singing along. Having lived through the fear of the remote reboot on nearly every platform for a while now, I'd say that it's pretty telling that I can let Software Update go on my remote machine nearly carefree.

It's been rock solid on my four week old Macbook Air.

MBPR 2014. Had a few issues (crashing, slow wake/boot, etc.) with 10.9 + 10.9.1, but most of these problems were resolved by 10.9.2. Haven't had a crash with 10.9.3 as of yet.

What I also realized was constantly upgrading from one version to another kept around some legacy files that messed up my OS a bit. A clean install might do the trick for those having issues.

That's the kind of advice that I constantly see in the Apple community boards for all kinds of issues.

Why is such a solution even acceptable? I had no issues before 10.9.3, but if I have to go through a clean install just to fix the issues, there's no way that Apple is going to keep me as their customer.

To be fair, "reinstall Windows" is oft-quoted as the best solution to a lot of Windows problems too.

That's true, but I paid ~3000 Euros for my Macbook (thats 4000 Dollars) and that's exactly why this isn't acceptable to me.

That argument makes no sense. It's a fundamental fact that technology isn't perfect and sometimes you need to fix things - regardless of the cost of the device...

And you would rather that your $4000 machine kept crashing rather then spending 30 minutes doing something that could fix it?

The problem is that this involves more than 30 minutes of my time. Setting up my development environment after the clean install will take half a day and I'm sure that I will forget to configure something in the process which will cost me even more time when I need the tool working.

I switched from Lenovo (which just worked without any problems) to Mac because I thought it's the same quality (it'll just work) with the bonus of allowing me to develop for iOS.

As it turns out, superficial things like the look and feel is better on the Mac, but it's nowhere near in terms of stability. I used my Lenovo for 3 years in work and at home as my only machine, whereas the Mac is in use for 1 day per week and I've already had more crashes on the Mac than I ever had on my old Lenovo.

I just don't have the time or the interest to babysit my laptop.

I've had no problems either. Early 2007 iMac, 2011 MBA, and 2012 mac mini all with 10.9.x (currently 10.9.3). I regularly switch out external displays on the iMac and MBA without issue. My coworker with an MBP (2012?) has had one kernel panic that I know about this week, however.

Yeah, my 2007 MBP is rock solid. It sounds like rMBP's with discreet graphics cards are having the most issues.

Same here. I'm running 10.9.3 on a 2012 Retina MBP, a 2013 Macbook Air, a 2011 iMac, and a few others with no problems.

In fact, it seems to have finally fixed the sleep wake failures I was occasionally having on the Macbook Air since Mavericks came out.

Ditto. Retina MBP and an older 11" Air and 10.9 has been solid. I never have to reboot either of them except to do some updates.

I've had more issues with hardware reliability with Apple products than with software issues.

No problems here with a 2012 rMBP 13". It appears the kernel panics are due to the swapping between integrated and discrete graphics. Since the 13" only has integrated, I've been experiencing no problems.

Same here, on iMac and MBA, both 10.9.3 as soon as it was out.

2013 MBA - zero issues. And this is on a corporate machine with Meraki, Sophos and Symantec Encryption Desktop running on it...

One kernel panic bug, and the system is "toxic" and "disease"? I'm not saying that this issue is acceptable, but I don't think those adjectives are really appropriate in describing it.

What's the Internet for if not hyperbole and over generalization?

This was 16 kernel panics, and was affecting myself and three coworkers who had upgraded to 10.9.3.

EDIT: ok, may be 16 panics of the same bug.

I suppose since most MAC stuff is described as "beautiful" and "stunning", which is also a somewhat extreme description for an email client or a file manager, then we have to have extreme negatives to match?

I've been a Mac user for 6 months (MBP), I hardly tax the machine but it has crashed a dozen times. Twice today, and five times this week in total.

Not since Windows 98 have I seen such an unstable OS. I've seen less BSOD in the last 15 years than I've seen kernel panics in the last week.

another mac user with the same problem. I've had 7 macbooks pros on various oses since 10.5 and osx is the first OS I ever just came to trust to run updates.

But now, 10.9.x is a piece of shit. It hardlocks all the time (2-3 times a week), seemingly around external displays. I'm really feeling burned at this point. These 6-7 locks in the past 3 weeks since I got a new laptop are probably 50% of my macos crashes since 2007. It's really really disappointing, and is making me question why the hell I pay this much money for a laptop. Apple needs to pull their thumb out of their ass and fix their fucking os.

This [1] seems to have helped a little bit, but seriously, fuck apple. I'm hunting for a way to downgrade my laptop to 10.8

Edit: to be clear, the really appalling bit here is how frequently these locks occur. This isn't a subtle bug; it's a symptom of apple being asshole cowboys. Given the frequency of these crashes, I believe it's impossible they used their laptop with an external monitor at all or their devs would have hit these too.

[1] https://discussions.apple.com/thread/5538012?start=0&tstart=...

I noticed that a system update installed itself yesterday at the end of the day. This morning I had two kernel panics (my first since getting the MBP a few months ago)... I'm also finding it seems tied to external monitors (both panics occurred when I tried to set an app as full screen on the external monitor).

I had to unplug and just stick with the retina display to get anything done.

I have seen increased lockups with 10.9.2. I replied because mine seem also to involve external displays (and/or Thunderbolt accessories). A possibly-related issue causing app lockups (Matlab) was caused by an old Java runtime interacting with multiple displays. Updating JRE fixed that. I have also gotten system-wide lockups on Thunderbolt plug-in (never used to happen with 10.8).

In my experience, people who've had exorbitant numbers of kernel panics on Macs have had tons of things installed that muck around with MacOS X internals. VirtualBox, old SIMBL hacks, manual fan controls, etc. Those are all notoriously unstable and can cause lots of kernel panics.

It seems more like a buggy OS that doesn't handle certain combinations of external monitors and GPUs.

Let's call it like it is: 10.9 needs some fixing. Plugging in an external monitor shouldn't be a big thing. I shouldn't have to tread on egg shells as far as what I do with such an expensive piece of equipment.

You mean all the thing I use to do my job?

I actually lag 4-6 months behind releases now on both my macbook and my iphone. While I used to love being an early adopter and downloading the latest release the night it came out. Over the past few years I've had so many stability issues that I can't afford the annoyance, frustrations, and hit to productivity. It definitely seems that the quality of apple os's out of the gate have more problems then they did in the past.

It may not be a process & quality question, but simply staffing with very highly experienced and smart people. Some analyze that the Windows kernel has that problem[1].

[1] http://blog.zorinaq.com/?e=74

Usually when I start getting a ton of kernel panics, I suspect hardware - bad memory or hard drive. Just had an iMac doing that last month and it was indeed the hard drive.

Anecdotal information is not particularly helpful but I have had almost zero 10.9.x kernel panic problems on any of our 11 machines except for one that had bad RAM and another that the hard drive was dying.

>VirtualBox . . . notoriously unstable

Is there virtualization software that is not unstable on OS X?

I've only had one kernel panic clearly linked to Parallels, but I still haven't made it to a month of uptime after over two years of using OS X, so it's quite possible that's it's causing non-obvious problems even with no VMs running.

I use VirtualBox every single day and never had a problem. One of my VMs has been running more or less continuously for more than 10 days.

VMWare fusion has been really good for me.

Some of the only kernel panics I have ever had we I think to do with VirtualBox's kernel module. VMWare feels faster too.

I've kept this same mbp up to date for 2 years and... It hasn't crashed once in that time.

I have to raise a counterpoint. I have, in my house, 3 Macs, two of them running 10.9, and the only one that ever panicked was the one that's not on 10.9. 10.9, at least here, is rock solid on both an aging MBP13 and on a 2 month old MBPr13. Both are regularly plugged to external displays, via DVI or VGA adapters and through AppleTVs/Airplay.

The only thing that bothers me is that, for some reason, all three have some difficulty reconnecting with the wifi after they wake up - they take more than 20 seconds to reconnect while the Linux laptops do it in less than 5.

Ditto on the wifi reconnection problems. I thought I had it narrowed down to the 5GHz band, but that seemed to be a different issue.

Also, I would say I've had relatively zero problems in recent years. Every now and again there will be a weird freeze, but so infrequently (i.e., a few times per year at most) that I don't blame the OS for it because I don't have any useful info.

Indeed. There is something very wrong with Brandon's Mac.

I have a mac mini at home and work primarily for development and neither of them have ever crashed. Anecdotes are fun.

I used to have a problem (twice) of extreme slowness in networking, but a reboot fixed the problem.

Please pastebin some of the panic logs.

My Windows PC never crashed like my Macbook does.

This is all anecdotal. Mac user for years and never had a crash. These types of comment provide very little to these threads.

I'm still developing on SL. That's one way of never having a crash or kernel panic...

It isn't just "anecdotal" when you happen to be the one who has to use a computer that crashes in such a way, regardless of who makes it, or what OS it's running.

Its data. The existence of crashes is evidence of something.

It may be data, but it is not a statistically significant amount of data. And there's an awful lot of selection bias in product complaints.

Crashes are not statistics. Code crashes or it doesn't. So something smells in Denmark.

Cmon. There's no such thing as a 'false positive' in code crashes. Its a binary thing - correct code doesn't do it. So something is wrong with an OS that crashes for somebody. Its got nothing to do with selection bias.

And if it crashes for certain people chronically, then there's a syndrome that's reproducible. Meaning it is likely to affect a lot of people in similar circumstances.

Crashes aren't new. OS X has had them for a long time, they're just less common than some other operating systems.

The question is whether you can go from "2 or 3 friends of mine said they have more crashes than before" to "10.9.3 is more prone to crashing than 10.9.2." For all we know, the entire sample is using some 3rd party utility that crashes under the new update, which is bad, but is also not something you can completely blame Apple for.

As a counter-anecdote, I've been running 10.9.x on my laptop since the first beta and haven't had a grey screen since beta 2 came out. On the other hand, my 1 year old Windows desktop running Windows 7 has had 4 BSODs this year.

Notably absent from the article, but spelled out in the comments, is that the author doesn't have an AppleID because he doesn't want to agree to Apple's terms, and therefore cannot acquire a kernel debug kit.

You're misquoting the author. He said that he doesn't have the time now to read through all the T&Cs.

That's fair. He's saying that he doesn't want to take the time to read the terms and therefore cannot in good conscience agree to them. A rare breed, to be sure, as commenter Dmitry Minkovsky points out.

> he doesn't have the time now to read through all the T&Cs.

This makes a lot of sense. IIRC, he's in the business of developing OSs and core OS-level technology. All he doesn't need is to agree with some silly terms and have Apple's legal on his back.

The safety of such agreements depend entirely on your line of work.

this doesn't make it sound any better unfortunately. perhaps i might be unpopular in saying this, but he should just click the fuck through. either you want to spend your time getting stuff done, or spend your time prevaricating about the bush.

Well, there's more to the story about the T&Cs, which I didn't get into because I was trying to write about panics... So I did try to create an Apple ID, which required I agree to "Apple Terms of Service", which links to http://www.apple.com/legal/. Ok, WHERE are the "Apple Terms of Service" on that page? Perhaps I'm being an idiot, but I don't see anything with that title. I browsed around for a while before giving up for now -- I've already lost time due to the panics, I'll get back to creating an account some other day.

Under the heading "Internet Services"

Ok, but which one? All of them? ... The tickbox says I'm agreeing to the "Apple Terms of Service" ... which means?

In any case, the spirit of my original comment was just to say that it's a notable omission from the article. I think that ansimionescu was right to point out the (perhaps subtle) difference between the author's actual explanation and my misquotation.

i might be being a bit grumpy, i guess it's estimable to take that kind of approach to agreeing to t&c's, but honestly on a case by case basis, you can quickly work out which sets of agreements you should spend time worrying about.

i think in his shoes i'd rather spend time debugging than peering over a somewhat self imposed roadblock.

This is exactly how freedoms are signed away.

I think the problem is in a different place.

Ask yourself this: If I post a piece of software on the internet, and on page 57 of the 132 page license agreement there is a provision that says you have to pay me $150,000 every time there is a lunar eclipse, what do you think as a policy matter that a court should do when there is a lunar eclipse and I file a lawsuit for non-payment in breach of contract against a million people who have used my app? Order them all to pay me $150,000 each? If so I think I have a new business model.

Attaching license terms to the use of software is problematic. When the terms are minor and unobjectionable then they're meaningless, because pretty much nobody is going to go through the trouble of litigation over the matter of someone lying about their age on a dating website, and what would be the remedy in that case anyway? Any terms significant enough to be worth enforcing in court are almost certain to be shocking or unreasonable, like requirements to pay money under unexpected circumstances or attempts to limit the rights the user has under copyright law.

More than that, we all know that 99% of users never read the damn things, and they never will. How can it be reasonable to enforce terms against someone who everybody admits never understood the terms?

To be clear, many/most software contracts limit the right of the user to do things like reverse engineer it, and some contracts also include rather controversial arbitration clauses.

Did you see this fun South Park episode?

https://en.wikipedia.org/wiki/Human_Centipad

Some people have felt that switching graphics modes is the source of Mavericks crashes. I had five crashes in several days last week. I went to the Energy Saver control panel, and turned off Automated Graphics Switching. Haven't had a crash in a week. Could easily be coincidence, but I thought I'd mention it here in case anyone else wants to try it. I'd be very interested in hearing whether it helps.

By the way, what happens on my machine is that the UI completely freezes. I'm not sure if that's what other commenters experience when they talk about "crashing".

Dammit. Finally updated last night.

No crashes or kernel panics since then, but I agree that switching graphics modes seems to cause most of my crashes. I get the "rainbow crash" where the UI freezes and garbage appears in stripes and blocks on the screen. I can still ssh into my MacBook and reboot it (or even restart the UI, I think).

I'm on 10.9.3 and do not see an "Automated Graphics Switching" option anywhere on the Energy Saver control panel... has this been renamed or removed?

Do you have a discrete GPU? Very few Mac portables have one anymore - obviously, the option doesn't apply if you have only CPU-integrated graphics.

I've experienced a lot of crashes since 10.9.3 on a late 2013 rMBP, mostly when sleeping/waking the device. Also, my WiFi is randomly disconnecting all the time since the update.

Before the update I never had any issues.

I love the hardware and the OS a lot, but I never had any crashes with my ugly Lenovo, so I'm currently thinking about switching back to Lenovo

> Also, my WiFi is randomly disconnecting all the time since the update.

I don't get the WiFi disconnects, but 10.9.3 has stopped being able to automatically connect to WiFi.

Wake up Macbook from sleep and find WiFi disconnected. Click icon, see only a couple of SSIDs, click the one I want, get the "unable to connect dialog". Click icon again and wait for refresh; now see a much longer list of SSIDs. Click the same one that it couldn't connect to before and it finally connects.

That's exactly how my currently Macbook behaves. I could write a short script to turn off/on the WiFi after login, but it's really annoying that you have to do that for such an expensive device.

Same here. Oddly, it always connects immediately at work, but fails in this same way at home, where I have an Airport Extreme router. You'd think if one combination was going to fail, it would be the one with two vendors.

EDIT: This is on 10.9.2.

I had one crash since 10.9.3 on my MBA 2011, and I have graphic glitches in Safari when displaying images since 10.9.1, I believe.

I've had only one kernel panic so far on my late 2013 rMBP 13. At least that's what I found in /Library/Logs/DiagnosticReports, and I've been through the 10.9.3 betas as well.

On the other hand I experienced random WiFi problems upon wakeup, had to do a couple of hard-reboots, suffered the rainbow static problem and other miscellaneous problems with external displays. Lately I've been experiencing some weird issues with the sound system: I hear music in my headphones but some notification sounds are sent to the speakers.

As a first-time Mac user (after a decade with Linux) I must say this is quite a frustrating and "meh" experience.

Love the screen and the battery life though.

Glad to hear I'm not alone, my colleagues rolls eyes when I rant about the problems I'm having.

Side Note: It seems to me that the most problematic models are Retina MacBook Pros.

I'm really surprised by all of the reports of crashing. I've had 3 different Macs over the last 7 years and I've had just one crash in all that time. I use my Mac anywhere from 8-16 hours per day and tax it pretty well. Before that I couldn't stay away from the BSOD on Windows, one of the reasons I finally made the switch.

My MBA locks up about once a week. Seems to be triggered by closing Firefox tabs with Flash content.

The only Windows BSOD I've seen since Win7 came out has been from a dodgy stick of RAM. A stick of RAM that then got passed around a few of us at work to ensure timely laptop upgrades.

Why am I getting down-voted for this?

The only reason I ever get bsods since windows xp is hardware problems. hence if you see bsod check out your box, do not blame windows. moreover, kernel panics on linux boxes are also most of the times hardware bound or you've broken something yourself.

Probably because 99% of people had not seen a single BSOD since Win7.

My external display just started going to rainbow static and back over and over again. Didn't think to blame the update till now. Also just had my first kernel panic in about a year.

I've seen that happen with Macbooks for years (the rainbow static), with both my mid 2012 retina MBP and the pre-retina one it replaced. Happens occasionally with my Dell monitor at home, and would happen with one of those knockoff IPS' that I used at my previous job pretty frequently. Unplugging and replugging the monitor will eventually fix it, but yeah, it's obnoxious.

Since installing 10.9.3, every time I wake my rMBP (mid-2012) up from sleep in the morning I get a full screen of static for 4-5 seconds. Subsequent attempts to reproduce fail, but the next morning it happens again. http://i.imgur.com/5dHFgN6.jpg

I also hadn't thought to blame the update for this until now (as I just have lost all faith in this particular model) but the timeline matches up well.

Edit: after re-reading, I think I know what you're referring to about the random static on the external display, which I've experienced over HDMI and Apple hasn't been able to debug (support case open for more than a few months now).

Experience a lot of crashes on my Macbook pro retina which is the latest model. Screen gets artifacts everytime the gpu swtiches from nvidia to iris pro or back. Coupled with the crashes that I now see on my Ipad air, I cant help get the feeling that things are starting to bitrot at Apple.

Apple have always had flaky products when it's something new. The first PPC macs, the first Intel Macs, OS X 1.0, and so on. Those of us that have been in the apple universe for a long time remember "don't use rev 1 of a product" was a mantra amongst Apple users since way back.

So, problems with the iPad Air are not particularly surprising. The shift to a 64 bit OS is a difficult thing to get perfect on a first go.

That said, they really should have nailed the discrete GPU kernel panics on rMBPs by now. That's been going in fir a couple of revs of the hardware and is unusual.

Anything with an A7 is a bit more understandable due to moving to 64 bit. 7.1 is already significantly better than 7.0 for 64 bit stability.

Safari crashes on a fairly frquent basis (> 10 times a day) even with the latest updates.

That's crazy. Is there a certain site that causes it more than others?

I've been dealing with Apple for months trying to get them to fix Mail in OS X Mavericks 10.9.2, which crashes on me about every 30 minutes, or sometimes more frequently. Their answer, which they refuse to put in writing, is that engineering is "aware of the problem."

Also, Mavericks broke Samba support completely. Apparently 10.9.3 doesn't fix it.

http://nelsonslog.wordpress.com/2014/01/15/mavericks-broke-s...

I agree with the comments about this being the least stable Mac OS ever. I'd put it up there with Windows ME. Multiple times per day I have to interrupt what I'm doing to deal with Apple-related bugs.

Strange, I use Samba every day at work, have had absolutely zero issues with it - this is across all versions of Mavericks, and previous versions of Mac OS...

Just checked my logs... not a single panic since 10.9.3

Interesting to hear people are having such problems. I'm on 10.9.3 and haven't had any crashes. Sometimes my cursor disappears for a few seconds or on top of certain windows, but nothing major.

I'm seriously unimpressed with 10.9 in general.

I bought a new MBP a few weeks back, and the hardware is awesome... the retina screen, 16GB RAM and 1TB SSD make it a joy to use.

But...

* I enabled Bluetooth, plugged in my Magic Mouse, and the next time it woke from sleep wireless connectivity had disappeared and has only come back sporadically since

* the machine occasionally 'loses' USB drives, complaining that I didn't eject them properly when they're still connected

* plugging in my Nexus phone via USB, or an HDMI cable, sometimes causes the keyboard to freeze, needing a hard reboot.

Rebooting twice within an hour isn't welcome when you're giving a lecture and have already met the student who loudly proclaims that Apple hardware is overpriced crap.

I do wonder whether Apple actually test new releases of their OS, or just leave it to us to do it for them.

And don't get me started about the usefulness of the "Genuis" Bar at my local Apple store...

I am wondering whether a downgrade to Mountain Lion may fix a few problems.

This honestly sounds like hardware failure. Bluetooth is very stable for me with a Magic Mouse.

Your USB and HDMI issues seem related. I'd wager a power issue.

Make the problem reproducible and they'll swap it over at the Genius Bar, who are extremely helpful if you're polite to them.

> the machine occasionally 'loses' USB drives, complaining that I didn't eject them properly when they're still connected

FWIW, I've had that issue several times and it's always turned out to be a hard drive which is on the verge of failing.

Random useless data point - my colleague on 10.9.3 plugged in a projector yesterday and had a kernel panic shortly thereafter.

I've had many, many crashes on 10.9.1-2 with my Mini.

Started when I connected my 22'' Dell over DP instead of HDMI with Apple-supplied DVI adapter, and added a second display – an Apple Cinema 20'' flipped to the side, to use it in vertical orientation.

I didn't wait for 10.9.3 to put that away. I just couldn't stand the way it would blank out the screens with a random color somewhere off my desktop; and hang – several times a day at times. Might have something to do with how crap our power line is (I've seen ceiling lights changing intensity throughout the day when this was happening).

Would be fun to check now with 10.9.3, my past experience tells me that the most whined about builds of OS X have worked out to be the most stable for me personally.

Most people here seem to refer to (r)MBPs only. I have a 2013 Fall MBA (a replacement I got for the failing SSD 2012 MBA) and the frequency of hardlock failures has been going up (currently once a day).

It also tends to happen a lot during web browsing more than anything else, I don't have any VMs, any external monitors, or any fucking-with-kernel things.

Ironically, it's been only 2 years since I started using OSX from Windows and Linux, and basically the amount my computers crashed hasn't changed a bit. I wondered why Apply fanboys kept raving about how crash-free their life was, since I wasn't seeing it.

I guess this is just what happens when an OS evolves to become an open enough platform for a lot of software to run on it, it just gets less stable.

Has anyone else had trouble with external thunderbolt<->dual dvi adapters in the 10.9.x series? I have a Mac at home that crashes the windowserver every time my external Monoprice IPS monitor is plugged in. It doesn't happen on a newer Retina mac at all.

At least you know the adapter is working if it crashes your computer. I am lucky to get the adapter to work properly and when it does, plugging in my Shimian crashes my MBP.

For what it's worth, my late 2011 MBA hasn't crashed once since I installed Mavericks. This article, though interesting, well research and well written is anecdata. Interesting anecdata, but anecdata non-the-less.

I'm seeing a lot of good reports for people with MBAs, and a lot of bad reports for people with MBPs. I think I can tell which ones the Apple devs use themselves.

I've had issues with freezes for a while pre 10.9.3 too. These obviously don't generate panic files so are more of a pain to try to debug.

Also I recently installed refind (http://www.rodsbooks.com/refind/) to dual-boot linux and have experienced a lot of issues as a result of that (separate to the aforementioned freezing issues.)

At least this guy has actual kernel panics :)

My late 2013 MBP 15 (16GB 1TB 2.6) randomly freezes 2-3 times per week, sometimes more. It seems to happen more frequently when under CPU load, but not exclusively. The mouse freezes, the fan slowly spins down, and it requires a hard reset.

I have two 1920x1080 external monitors (hdmi and thunderbolt) which seems to be a commonly reported configuration with crashes.

It came with Mavericks and has done this ever since I got it, under all versions of Mavericks

No kernel panics here, but I get weird behaviour with external displays.

Sometimes when I hook up my external displays, some of the icons on one of the menu bars will disappear. It's very strange. I can fix it by going into the display settings and changing which one is primary.

I've got a recent rMBP with 2 x 1920x1200 monitors attached, one via a MiniDP -> DVI adapter, the other via as MiniDP -> Dual Link DVI adapter.

If we're going to collect anecdotes then I'll note my 2013 MBP has been running continually for over a year without a panic. I have a 2006 MB and 2008 MB (both still running) each having panicked 2-3 times over their lifetime. I'm a developer, my son is a musician and my wife is a graphics design artist so there's been lots of software thrown at these machines.

And here's mine: My first Mac was the G3, I'm now on a 2013 11" Macbook Air (10.9.2) and I have never experienced a kernel panic ever, I couldn't tell you what a kernel panic looks like. The only time I ever have an issue is if I try waking my computer immediately after entering sleep when attached to an external display – it doesn't happen and inevitably requires a restart.

    Current uptime: 23:05  up 3 days,  8:36, 2 users, load averages: 1.67 1.56 1.67

I have never experienced unreliability, and it's a pleasure to use and carry. I tried Ubuntu on a HP Laptop I had around, and I didn't like it one bit, it felt slow and the UI felt clunky. Even when using SSH to access the machine it took around 5-6 seconds for the bash prompt to appear, compared to an almost instantaneous Debian prompt.

I got a couple of kernel panics on upgrading to 10.9.3 but then Nvidia released new drivers for it and i haven't had a single kernel panic since switching to them.

Although - one port on the graphicscard won't wake up if a connected screen goes to sleep. Nvidia had that problem on the first released 10.9.2 driver, but fixed it about a day later.

Could you try checking if the Nvidia supplied drivers helps?

I've had no kernel panics on 10.9.3 ... but this Haswell rMBP 13" periodically froze after password entry when screen locked. And froze so badly that switching user was no way out: I had to hard-reboot. Since disabling the screen lock on inactivity it's been okay, but that's not a terribly secure or reassuring solution.

The only weird problems I’ve had on 10.9 at all have been with the WindowServer locking up when coming back from sleep or lock. This is on a late 2013 MacBook Pro.

After a ton of messing around and talking to Apple Support, we found that trashing the WindowServer plists that are stored in multiple locations seemed to sort it out… so far

Weird. Fr me, I can't remember my MacBook Pro hasn't having a kernel panic in the last six months, and at one was one of the only ones I've ever had. I recall that it was something to do with a VirtualBox kernel module or something, so I upgraded VirtualBox and have had no problems since.

My 2008 Mac Pro has been rock stable since 10.9.2 (before that, just the 'mail, are you drunk?' issues).

Sounds a lot like the problem is graphics switching - everyone with issues here has two graphics adapters.

MBP 15 inch (mid 2012) gets ultra hot. I have never seen a laptop become this hot. It also has various problems getting back from sleep and low battery. So very annoying.

Sadly, I am sort of forced to use it everyday.

I have a late 2013 15 inch MBP that regularly reaches 100 degrees C when running Adobe Flash. The only other time the machine has exceeded 70 is when I tried running Steam games at 2880x1800.

The hardware simply shouldn't get that hot, no matter how poorly optimized the software might be.

I've had terrible issues with my rMBP 13" (freezing, kernel panics) until, 1 year after I bought this machine, the SSD finally went out. Got it replaced and no issues so far.

Older OSX updates have done the same thing to my similarly older Macs. I wish Apple would be more diligent with their testing - especially when it comes to video.

10.9.3 has been nothing but absolutely stable for me, and I do a ton of development tools installed and do loads of tweaking and customization.

i'm sure that the stock Apple support answer would be to backup, reformat and restore.

This would probably have fixed his problem since the backtrace he finally Obtained fingered the file system, which is probably corrupt. (not that this excuses the panic!)

This is further evidenced by his "helpdesk" being unable to repair his computer by physically swapping drives.

Yes, this is a good theory, and consistent with the decoded stack trace we have. Although, consider that three other coworkers are suffering the same panics after upgrading to 10.9.3. Have we all got the same corruption? I suppose that could be possible, if the corruption was caused by the same bug.

I have the OS X 10.9.3 on a Macbook Pro , Mid 2009 and there is absolutely no problem with it. Is it on newer macs?

not just 10.9.3, entire 10.9.x is the crappiest OSX of all times ! So many stability issues.

Well, the 10.9.x Finder is super buggy since the start, perhaps 10.9.3 just made it worse.

No problems over here at all.

We are having problems, when using multiple displays connected to the iMac.

Disabling hibernate and fighting with power napping has solved most of my crashing problems. I think 10.9 is a mess when it comes to trying to save me power.

I also still disable noatime, journaling, swap space, and sudden motion sensors on any computer with an SSD; even if its unnecessary.

If you disabled journaling and swap on your machine, you're quite far on the way to data loss.

Lots of crashes here too. In the graphics driver.

“It happens to me therefore it is toxic!!!!!!"

Anything that fucks up your workflow is toxic (very bad, unpleasant, or harmful).

No, it's happening to myself and three other co-workers.

I use both thunderbolt and my HDMI out on my rMBP daily.

This guy is a bafoon complaining about being unable to fix a problem by not accepting T&C, but he's using OS X?

I'm not denying his problem, but he is certainly cause for it being a lot worse for him than it really is.

Aren't bugs in pre-release software expected? He makes it sound like there are tons of issues, when it looks like this is just same one bug related to external display handling.

This isn't prerelease software. This is the current version of the OS that Apple will prompt you to upgrade to automatically.

10.9.3 is not pre-release software.

Reply on news.ycombinator.com