[eleanorb]

Almost every morning, Safari quits unexpectedly. This seems to happen only once a day and it doesn't matter whether I've used it since opening the computer or not. (I leave apps open overnight and just open the lid to start working again.) Here are a couple of days' console outputs around this event:

2015-08-16 7:21:28.359 AM lsboxd[285]: @AE relay 4755524c:4755524c
2015-08-16 7:29:12.336 AM com.apple.launchd[1]: (com.apple.XType.FontHelper[2321]) Exited: Killed: 9
2015-08-16 7:29:12.000 AM kernel[0]: memorystatus_thread: idle exiting pid 2321 [XType.FontHelper]
2015-08-16 7:29:12.638 AM com.apple.launchd.peruser.502[142]: (com.apple.pbs[2310]) Exited: Killed: 9
2015-08-16 7:29:12.000 AM kernel[0]: memorystatus_thread: idle exiting pid 2310 [pbs]
2015-08-16 7:30:54.856 AM com.apple.launchd.peruser.502[142]: ([0x0-0x27027].com.apple.Safari[316]) Exited: Killed: 9
2015-08-16 7:31:14.702 AM com.apple.launchd[1]: (com.apple.coremedia.videodecoder[1872]) Exit timeout elapsed (20 seconds). Killing
2015-08-16 7:31:25.534 AM com.apple.usbmuxd[21]: stopping.
2015-08-16 7:31:25.784 AM com.apple.usbmuxd[2356]: usbmuxd-345.70 on Jan 26 2015 at 16:55:11, running 64 bit
2015-08-16 7:34:04.188 AM WindowServer[84]: Created shield window 0x333 for display 0x04272780

2015-08-19 6:54:55.976 AM com.apple.launchd.peruser.502[142]: ([0x0-0xa7ba7b].com.apple.inputmethod.ironwood[5854]) Exited: Killed: 9
2015-08-19 6:54:56.000 AM kernel[0]: memorystatus_thread: idle exiting pid 5854 [DictationIM]
2015-08-19 6:54:56.595 AM com.apple.launchd.peruser.502[142]: (com.apple.CalendarAgent[5792]) Exited: Killed: 9
2015-08-19 6:54:56.000 AM kernel[0]: memorystatus_thread: idle exiting pid 5792 [CalendarAgent]
2015-08-19 6:54:57.415 AM com.apple.launchd[1]: (com.apple.sleepservicesd[5781]) Exited: Killed: 9
2015-08-19 6:54:57.000 AM kernel[0]: memorystatus_thread: idle exiting pid 5781 [SleepServicesD]
2015-08-19 6:54:58.000 AM kernel[0]: memorystatus_thread: idle exiting pid 5786 [cfprefsd]
2015-08-19 6:54:59.693 AM locationd[6000]: NOTICE,Location icon should now be in state 0
2015-08-19 6:54:59.879 AM com.apple.time[145]: Interval maximum value is 946100000 seconds (specified value: 9223372036854775807).
2015-08-19 6:56:43.388 AM com.apple.launchd.peruser.502[142]: ([0x0-0x938938].com.apple.Safari[5308]) Exited: Killed: 9
2015-08-19 6:57:16.000 AM kernel[0]: (default pager): [KERNEL]: default_pager_backing_store_monitor - send LO_WAT_ALERT
2015-08-19 6:57:21.000 AM kernel[0]: macx_swapoff SUCCESS
2015-08-19 6:59:47.924 AM WindowServer[84]: Created shield window 0xb9f for display 0x04272780
2015-08-19 6:59:47.924 AM WindowServer[84]: device_generate_desktop_screenshot: authw 0x0(0), shield 0x0(0)
2015-08-19 6:59:47.935 AM WindowServer[84]: device_generate_lock_screen_screenshot: authw 0x0(0), shield 0x0(0)

Any suggestions? This is Safari 6.2.7 on a 2009 MacBookPro using OS 10.8.5

Thanks, Eleanor

[I-ku-u]

I presume you mean a MacBook Pro since MacPro's don't have a lid. (-:

Anywho, the logs indicate Safari is being deliberately killed, forcefully, by launchd (which has every right to do so, btw, since it controls what runs to maintain the OS' overall functioning). launchd is also killing other processes, and the logs show it's not consistent which those are. But that means that Safari itself is not likely to be the problem, it just gets caught in the crossfire more often than not.

It looks like you've posted only some of the logs, given the three day gap between the events. However, if those are indeed all the log entries, then it looks like there's a system problem related to putting the laptop to sleep (I suspect It's a bad sign when even sleepservicesd gets killed).

How long has it been since you restarted the laptop? If it's been a few months, such inconsistent behavior wouldn't be surprising. Launchd could also be reacting to a hardware issue - which you might detect if you run hardware diagnostics.

Those are my only clues to share. Good luck and I hope I've helped.

[eleanorb]

Thanks for your attention to my post. I generally restart my computer every week if it hasn't had some problem that needed a shutdown before then. The logs above are complete for each event — that is I picked a point before and after the event and then included every log message between the two points.

Safari didn't get killed this morning, but a bunch of other stuff did,. I tallied the number of times each got killed in that period — you might be interested.

22 hours on 8/20-21 (11am to 9 am)

(com.apple.sleepservicesd[8005]) 81
(com.apple.inputmethod.ironwood[615]) 45
(com.apple.tccd[7972]) 20
(com.apple.CalendarAgent[7971]) 18
(com.apple.hiservices-xpcservice[7939]) 9
(com.apple.xpcd.[7940]) 9
(com.apple.pbs[7935]) 7
(com.apple.cfprefsd.xpc.agent[8003]) 6
(com.apple.printtool.agent[7941]) 6
(com.apple.accountsd[7965]) 5
(com.apple.cfprefsd.xpc.daemon[7993]) 5
(com.apple.coreservices.appleid.authentication[7490]) 2
(com.apple.tccd[7302]) 1
(com.fitbit.GalileoMenubar[225]) 1
(com.apple.ImageCaptureExtension2[308]) 1
(com.apple.audio.ComponentHelper[788]) 1
(com.apple.audio.SandboxHelper[787]) 1
(com.apple.coremedia.videodecoder[4801]) 1
(com.apple.FolderActions.enabled[196]) 1
(com.apple.mdworker.shared[8125]) 1
(com.apple.qtkitserver[786]) 1
(com.apple.quicklook[8109]) 1
(com.apple.security.pboxd[7928]) 1
(com.apple.UserEventAgent-Aqua[145]) 1
(com.apple.XType.FontHelper[793]) 1
(com.dervishsoftware.MailHubUpdater[195]) 1

It seems (to me) that sleepservices gets killed a lot when the computer as a whole is not being put to sleep. As in the previous logs posted, sometimes the killer is com.apple.launchd and sometimes com.apple.launchd.peruser — do you see any significance in that?

What hardware diagnostics do you have in mind? I did a disk verify and all was well.

Thanks,
Eleanor

[FrankMtl]

Can you run your Mac in safe mode for a day or 2? Reboot holding the Shift key until you're back at the desktop. Note the time you rebooted and keep Console running.

[eleanorb]

Yes, I could do that. I'll wait a bit, however. Yesterday I upgraded Safari to 6.2.8 and also I am keeping Firefox off (I believe it was on for much of the previous time), so I'll wait and see if the kill-Safari bug happens again. If not, then one of those things was doing it. If it happens again, then I'll try your safe mode suggestion. Thanks! Eleanor

[I-ku-u]

This page from Apple ought to tell you how to do the hardware diagnostics I mentioned:
https://support.apple.com/en-us/HT201257

And there likely is no meaningful difference between launchd and launchd.peruser being the killer, since the distinction primarily exists to separate between processes owned by you or the system.

[I-ku-u]

I just had another thought, which could also be very relevant. How often do you see the LO_WAT_ALERT in console messages? When I first saw it I thought it was power related, but then I realized it's from the pager.

The message itself doesn't appear to be an error, as it seems to be normal when your mac creates a new swap file. Creating a swap file isn't an error either, but you might get this behavior if the boot disk is nearly full, lacks room for a new swapfile, and some other relatively critical system process needs the RAM.

[eleanorb]

I can't get the AHD to load - when I follow the instructions I don't see the icon mentioned, and then eventually I get the usual signin display. I believe my computer was delivered with Snow Leopard, but the install disk I have is only 1, no Applications Install Disc 2, as mentioned in the Apple instructions. I upgraded to 10.8 (with considerable pain), and now am 10.8.5. A cursory search on the internet and in MacNN yields no solution. Do you have any suggestions?
Eleanor

[I-ku-u]

If you lack the necessary disc, then the only ways I can imagine to try the hardware test is to contact Apple, or find someone else near you with the needed disc.

[eleanorb]

I don't understand why a hardware test is so chancey - varies by computer, by system shipped with, etc. Not exactly a standard remedy. I'm surprised Apple even pretends to support it, and not surprised that in 20 years running Apple computers, I've never before heard of it! Eleanor, in frustration

[I-ku-u]

Originally Posted by eleanorb 

I don't understand why a hardware test is so chancey - varies by computer, by system shipped with, etc. Not exactly a standard remedy. I'm surprised Apple even pretends to support it, and not surprised that in 20 years running Apple computers, I've never before heard of it! Eleanor, in frustration

It varies because the technology used to build the computers has varied. Apple has worked to keep it consistent, but they'll change it when they can make it easier. It also usually isn't necessary as most hardware issues will prevent the computer from running in any useful manner. Most, but not all.

Since usually a software issue is the cause of troubles, it makes sense to check the software first. But it's a question of effort to return - the easier it is to check the hardware, the earlier it makes sense to run the test. I wouldn't stress just now about running the hardware test unless you find a way to do it conveniently.

Have you looked into my other thought yet, about your RAM usage and free disk space on your boot disk?