- Joined
- Jun 28, 2015
- Messages
- 26
- Motherboard
- Gigabyte Z170X-UD5
- CPU
- i5-6600K
- Graphics
- HD530 + GTX 1060
- Mac
- Classic Mac
- Mobile Phone
Hi all!
______
Meta-Comments
My machine is now working, but the root question of why the arbitrary freezing (without anything in the crash log) — including in the Sierra installer and in an installation complete but not set up (e.g. user account) (not to mention a re-installed El Capitan under Safe Boot) — remains unsolved.
The question thus is…
Did anyone else with this (or any other) motherboard get arbitrary freezing during installation?
My new Sierra installation is on a different drive; the original El Capitan, that would freeze arbitrarily, was on the same drive as Windows 7 RAID (defunct), Windows 7 and Ubuntu Linux.
I have edited out the comments about Blu-Ray. (They were somewhat vague anyway, since I did not then have access to the details.) The bottom line is that the Blu-Ray DRM system is set up to not work unless you pay — not only for the disc but also for the player ( https://www.howtogeek.com/240487/how-to-play-dvds-and-blu-rays-on-linux/ ). I installed a free version of one commercial player (software), and it worked perfectly, albeit with a large text banner in the centre of the screen all the time.
______
— If you have just joined us, you can jump straight to Updates — the last material. […Although there are questions in the thus-skipped section.] —
************ BTW: I ask a lot of incidental questions herein, but the core question is that of what is wrong with my machine and how to fix it.
I have not taken notes detailed enough to be adequate for it, but here is the situation from memory. Sorry if this seems excessively detailed, but it is that kind of problem.
This should be irrelevant (I imagine), but:
Motherboard — Gigabyte GA-Z170X-UD5 (*not* TH)
CPU — i5 6600K
RAM — as per text; currently G.Skill F4-2400C15D-16GVR, running without XMP (or whatever it is) at 2133 [thanks UtterDisbelief].
Edit<
Storage — an excessive collection of 3TB HDDs. My boot volume is the same for Windows, Linux and MacOS — all manually assembled. [I should be able to answer basic questions about configuring Windows for multiple boots; the secret is: duplicate, not create new entries.] That is, I have all my bootable partitions on the same drive — a 3TB Western Digital Blue. (I somehow ended up with a spare of the same type; at one point recently I was trying to clone my boot drive onto that.)
>
Let me note at the outset that the machine is perfectly happy at the moment in Windows or Linux; it is only in the Hackintosh that it freezes. Also, it still freezes under Safe Boot.
<This part should (*shoouuuld!*) be irrelevant. (I include it for completeness’s sake.)>
I originally bought some fast, large RAM — 32GB@3200. [This was a snap decision that I regret.] (It was F4-3200C16D-32GVK, but that does not matter). The point here is that it was not listed against my motherboard — neither on the site of the motherboard manufacturer nor of the RAM manufacturer (being Gigabyte and G.Skill, respectively)… but I did find someone on the internet who had that combination and reported that it worked fine.
As I was setting up my Hackintosh, it did not faze me that it crashed here and there. Finally, however, it became apparent that it was crashing — freezing, particularly — once every one to two weeks, arbitrarily — including when it was ostensibly idle.
I found out (rightly or wrongly) that this is a RAM issue. I bought TechTool Pro and ran it, and it reported a stuck address. (Note that this runs under the Hackintosh OS. That is to say that, at this point, I had not actually put in the RAM settings in config.plist; I would have gotten around to that much earlier under different contingencies. Of course, I should have realised that running the check under MacOS was therefore not a good idea. I did actually think of this, but I thought that the program would talk directly to the RAM. (Maybe it does!) )
I took it back to TechBuy, and they duly returned it to G.Skill. I was told that it would take about three weeks, so (for reasons I can not recall offhand) I felt it necessary to buy some RAM for the interim. [Edit: … and I installed this.] Actually, I had the replacement RAM back in about three days.
I found out later that TechBuy had written, “failed memtest”, on the paperwork; I now infer that G.Skill took my word for it that the RAM was indeed faulty; I was actually hoping that they would check it themselves. My guess, now, is that it was fine — just incompatible.
It turns out that you can download the excellent “memtest” for free, and it boots itself and everything. All my later checks were with this.
I did not bother to actually replace it [i.e. install the replacement instance of the original RAM] in the machine, and ultimately it was about six months before I did this. (Everything has been a nightmare for this whole period.)
The point is that I had the substitute RAM (F4-3200C16D-16GVKB) installed in the machine for six months, and it was fine as I recall (or maybe it only froze every four weeks or so)… and this substitute RAM is also *not* listed as tested to be compatible — a stupid and embarrassing mistake on my part. [The various RAM is “original” and “replacement” for the 32GB@3200, and “substitute” for the interim substitute that ran for six months (all G.Skill).] [By the way… if anyone knows why RAM manufacturers have to test their RAM against individual motherboards to see if it actually works, I would be very curious to hear how that makes sense. (Do they genuinely not know how?!!) ]
Recently, then, I have installed the replacement RAM in the machine, and it was not okay — I ran memtest, and (from memory) it reported that it was fine on the first run, but on the second it reported a collection of errors at the top end, all on the same bit in a 32-bit word (or whatever it is). I have been told that that is characteristic of incompatibility. I think the machine was crashing about once a week at that stage.
I then tried the substitute RAM; I can not recall the details, but I would have run memtest on it, and the outcome was that I raced out and bought some (!!certified!!) RAM. (If memory serves, it was crashing, and I ran memtest and it reported no errors, and finally I ran it again and it did report errors.)
</This part should be irrelevant.>
The outcome of the foregoing is that I now have compatible RAM in my machine, and I have run memtest on it twice to make sure.
So…
This is where I would have liked to have comprehensive notes. Around this time [i.e. during the ultimate sorting out of the foregoing RAM issues ((or believed I had?)) ], the freezing went from under a week to under 24 hours. In hindsight, something must have changed in the software side, during this period; I did not realise at the time that this represented a systematic change. [See also “Notes”, below.]
I am thinking that what has happened is that I have fixed a freeze caused by incompatible RAM, and coincidentally developed a software problem.
There are only three candidates that I am aware of — my best guess is the third. This is *not* to say that I am in the slightest ruling out some other cause.
One is that I tried hard at some point to get software that would play Blu-Rays. (I had found out previously that this is not a trivial exercise.) I have installed a variety of players to try them. (I have Kaspersky anti-virus software.) I am sure that I would not do this when my computer was crashing every 24 hours. Conversely, I have a vague memory that it was a couple of weeks or more earlier, that I was fooling around with Blu-Ray players — the point being that that would rule that out as a cause. Still a possibility.
The second candidate is the audio. Again as described below, I came to suspect that this might be the source of the problem, and had trouble trying to fix it. At this point(as I write), the AppleHDA.kext is in the Trash, and everything else should be normal for a Voodoo installation. [My understanding is that Voodoo has a disabler for the AppleHDA, which should imply that Voodoo would be working.]
The third candidate is double-barrelled, as I understand it. I found out — through doing a safe boot (which checks the hard discs) — that I had corruption on my hard disc drive. Initially, it did its thing fixing it, and then reported that it could not fix it, and turned off the machine. When I turned it on and did another safe boot, it tried and said that it had fixed it. I know enough about these things to know that it is not fixed — that the first run made some changes that locked in an error.
The point is that it was already in (what I am taking to be) the new software crash era, before this “fix”ing, and that after this it was much worse — it went from crashing within 24 hours to crashing during the startup phase — the part where it has the desktop showing, and is pretending that it is up and running, but actually takes a few minutes to finish starting everything up. It does not always crash during startup; it ranges between that and five or ten minutes later. (The best is about 5 hours, but that is a massive outlier.)
This brings us to the recent past.
I came to suspect the audio — Voodoo 2.8.8 — on the basis that… (since I reinstalled it ages ago) it always comes up with a message that it could not load the Voodoo kext (plus a few coincidences). There were 2 instances of the kext — one in S/L/E* and one in L/E. From memory, offhand… I deleted the one in S/L/E, and later/soon came back and moved the other one from L/E to S/L/E. (I remember being surprised that the sound suddenly came on while the machine was running; that would be this; (at least on my machine) it has to be in S/L/E.) Early on in this episode, I reinstalled El Capitan and the Multibeast set. Particularly, I tried installing the Series 100/ALC1150 version; it does not work on my machine, but I thought that perhaps it [the machine] would not crash continually.
[ * <hard disc>/System/Library/Extensions ]
Finally, I trashed AppleHDA.kext from L/E. (In hindsight, this was very foolish, but at the time I was thinking that maybe it was the AppleHDA.kext that was causing the problem. [As well as I could tell, the backup and the original were exactly the same — which I thought was rather strange; there is no point carefully replacing the original if the substitute is no different (not to mention having a substitute in the first place).] )
I think my problems have worsened significantly from that time.
My project at the moment is to do a low-level backup of the machine as it is, and swap in a copy of the partition from a while ago (plus all the attendant work stuffing around with my work that is thereby deleted), but that endeavour, too, seems to be cursed.
In summary, then, my Hackintosh freezes under safe boot, after reinstalling the OS and the MultiBeast stuff (but the hardware is fine under another OS).
Any help would be deeply appreciated. (It would be *far* less of a hassle to fix one software problem on the current setup, than to roll back to a backup from several months ago and then have to transfer back in all my mail and documents, plus all the work I have done on the machine in the interim.)
_________
[Context: I [will] have backups of the partition as it is now, and as it was several months ago.] _ One concrete question… . Assuming that it is something related to either • audio drivers or • software installed by a Blu-Ray player… given that I have a Clonezilla backup of the partition from before the (new) freezing started, and assuming that I can mount this and (Linux forbid!) actually copy the files… is there some group of files that I could copy, that should get rid of the problem, such as to save me reinstalling/transferring all my email and browser history and bookmarks and documents and what-have-you, please? (This would save me a *lot* of hassle and time!) For instance, if I were to copy all and only • System(/Library) and • Library, then that would delete all my mail, bookmarks and so on [or possibly some of this is stored in my User folder?], but if I were to copy only (say) S/L/E… would that do it? [p.s. I do not understand why copying the AppleHDA.kext back does not fix it; it there any form of copy, short of a partition replacement, that will work?]
Speaking of which… I have posted a question about this online, but I might as well ask it here as well. Linux flatly refuses to have anything to do with files such as those copied from a Macintosh into a backup, on the grounds that the permissions prohibit it. (The source has the wrong owner, and the target is read-only.) Ignoring the issue of having to learn command line flags for every little thing, and wondering if one has made some terminal mistake… the “solution” seems to be to actually change the ownership and permissions of the files. (a) If I do this to system files, would a permissions repair put it all right again? Is there (say) a hacked version of Linux that is not obsessed with checking permissions? (Can I set up a user account called “root”, and everything will just work?) On the Windows front, the problem is that the geniuses at Microsoft put a limit of (about) 256 characters on path designations. Is there a hack out there to fix that little oversight? [Later: apparently not.] (Apparently there is in Windows 10, but I am using 7. I am just trying to see if MacDrive will prevent this. I have Paragon’s HFS+ thinggy (which does not).)
Incidentally. Is it theoretically possible to put together a (presumably large (at the time of writing)) USB stick with a bootable HackOS (including the Finder) plus the ability to (a) copy Dd and Partclone *partitions* and (b) the ability to copy *files* from said? If so, how far is this from immediately practicable, please? (I have just been searching; I found this https://www.insanelymac.com/forum/t...ow-to-make-a-bootable-leopard-on-a-usb-stick/ (from 2008), and then stopped.)
If so, can it be given the ability to read dd and partclone containers — this being the format of my backups (done with/by Clonezilla)?
Notes
As far as I can tell, there is nothing in the Console logs, ever. (There are excessive wake-ups going on, but I take it that is not about sleep (as in low power), and that it is pretty much normal.)
I should note that I have a Skylake processor, and I am only on El Capitan, and my machine setting is “17,1”, so my operating system does not know about the processor. I tried once telling it that it was “14.2” (without making any other modifications), but it still crashed a moment after booting.
I have tried upgrading to Sierra, but, in most of my attempts, the machine has crashed before it finishes the 1-minute setup, and in the one case in which I got to the reboot point, the {continue with installation} boot option was not there. (I tried booting normally, but that did not work.)
I have been trying to get Blu-Ray movies working on the machine, and I have installed all the usual suspects, including one that requires Java Runtime (SE) 6. Perhaps unwisely… I found a terminal script, obviously written by someone who knew what they were doing, that purported to get VLC working properly… that (among other things) changed the permissions of some pertinent files.
Informing the third candidate, above: I have been amalgamating instances of my stuff (applications, utilities and documents) from about four sources — i.e. various past machines — preparatorily to copying elements (as is indicated) of this amalgamation onto the machine in question. The point is that I may have inadvertently copied something corrupt onto my gleaming new machine — for what it is worth.
Updates
20180501
Trying to boot and shutdown cleanly in MacOS, so that Clonezilla will not balk at copying my partition. Instead, I got many, many instances (most the same entry, oddly enough), of:
hfs_swap_BTNode: record #38 invalid offset (0x0000)
hfs: node=16199 fileID=4 volume={My boot drive} device=/dev/disk5s12
hfs_mark_inconsistent: Runtime corruption detected on {my boot drive}, fsck will be forced on next mount.
Also:
<Emergency> Boot task failed: fsck-safe
By the way: I ended up with 2 instances of a HD Recovery partition. That is not going to cause any problems, is it?
I guess my next move is to reinstall El Capitan again, and try again to get a clean (boot and) shutdown.
That is… it is looking more and more as though there is something actively screwing up my disc (that is, this volume), and that the best alternative is to nuke it with a previous image and go through all the hassle of updating that.
I do have some corrupt files copied from failing previous machines, but they are on different volumes. I have been copying files onto the new machine, but there should not be any corrupt files among those (surely!?). …And even if there were, they would not actively damage the disc after it was repaired.
Thank you all for taking the trouble to read this!
… I am developing a strong-ish suspicion that the Linux endeavours to repair my hard drive (preparatory to Clonezilla backing it up) are what is causing my immediate problem (or are part of the problem or whatever). Last time around, when I was trying to do a backup with Clonezilla, I told it to do repairs as necessary [noting that there *were* errors], and I remember that it passed the first three layers without comment… and when I next looked at it with Disc Utility, *that* failed absolutely to make any sense out of it, and gave up. (I would have thought that if HFSProgs was bad, I would have stumbled across mention of it several times in the past.)
… Or maybe it does not understand Journalling, and it is all my fault for trying to use it to do repairs. Trying Dd.
<Edit/Note:
In my desperation, I accidentally allowed a Linux program to do a repair of an HFS+ partition with Journalling still on, thereby making it much worse… noting that, now, I would not choose to do this at all (even with Journalling turned off). I am almost sure that my hard disc was already corrupt, at this point (which would be why the Linux program wanted to fix it). It would seem that Journalled HFS+ is *not* merely HFS+, plus historical notes — such that one could work with it as though the notes were not there.>
[For what it is worth, I note in my defence that GParted explicitly notes that there are unapplied journal entries, and offers to Fix this (and apparently fails to without bothering to report any errors); this seems to me to be inconsistent with it not supporting journalling.]
/ Edit/Note>
20180503
Okay; this isn’t funny any more (not that it was before!). I have just cloned my hard drive onto another one (so that I have some hope of recovering my files from the broken file-system), and replaced the contents of the HackOS partition with a partition backup from a few months ago. [For avid readers: that is from the period when I was still thinking (correctly, I believe (but not absolutely) ) that my problems were caused by incompatible RAM.]
It booted fine… Disk Utility said the partition was fine… and it still froze after 1hr04min. The RAM is fine. (I have checked it twice.) The only thing apparently wrong is that it said that it could not run the VoodooHDA.kext — the reason for that being that I have ended up with 2 copies — one in S/L/E and one in L/E. Arguably, if it refused to load it, it can not be causing any problems.
! If I failed to re-install AppleHDA.kext (or whatever the Apple sound thinggy is), but instead merely copied the file back in (not realising that that was not adequate), would that cause this problem, please, anyone? (Just out of curiosity: if that obtains, can I get the machine running by deleting the thing entirely?) I suppose that would explain why the installed copy was exactly the same as the backup copy — that is, that I had previously copied it back after trying ALC1150 and then changing to Voodoo. If I install ALC1150, would that disable AppleHDA and fix everything — if this is indeed the problem, please?
Again: the machine is perfectly happy at the moment in Windows or Linux; it is only in the Hackintosh that it freezes. Also, it still freezes under Safe Boot.
Any help greatly appreciated; thank you in advance.
20180512
(After a struggle) I managed to get Sierra running; it has run without freezing for over 24 hours, at least twice. {Emoticon for laughing hysterically then fainting.} (I get a sense of glee and deep satisfaction from moving the mouse pointer around. Sad!)
The really intriguing thing is that it freezes (as above) while the Sierra installer is running (both while booted from the USB (e.g. while installing) and while booted from and still setting up the new OS). Apparently there is something about completing a (Sierra) installation (and rebooting) that takes it from freezing to not. [Just repeating: there is never anything in the logs (unless I am not reading them correctly).]
It was indicated, ultimately… that there was some corruption in my El Capitan installation… but that does not work for booting from a Sierra installation USB. (Supporting the foregoing tenet is the fact that it was (mostly?) okay for months, several months ago.) Unless this happens to everyone, the indicated conclusion is that there is something in my hardware that is fine under Linux, Windows 7 and Sierra, but not under El Capitan nor the Sierra boot environment. … Something entirely unique to my (particular instance of my) hardware. Even if I imagine that (for instance) my motherboard is faulty, it is hard to make sense of it all — not that I am an expert on hardware fault symptoms. It is perhaps more likely that there is some anomaly in the Clover settings on the installation USB (plus coincidental corruption of my El Capitan install)… but then there remains the question of why this affects only my machine (GA-Z170X-UD5, i5 6600K).
? Did anyone else with this (or any other) motherboard get arbitrary freezing during installation?
I shall probably never know what it was. (It would be interesting to do a clean installation of El Capitan, but that is certainly not a priority right now. I expect it would work fine… .)
______
Meta-Comments
My machine is now working, but the root question of why the arbitrary freezing (without anything in the crash log) — including in the Sierra installer and in an installation complete but not set up (e.g. user account) (not to mention a re-installed El Capitan under Safe Boot) — remains unsolved.
The question thus is…
Did anyone else with this (or any other) motherboard get arbitrary freezing during installation?
My new Sierra installation is on a different drive; the original El Capitan, that would freeze arbitrarily, was on the same drive as Windows 7 RAID (defunct), Windows 7 and Ubuntu Linux.
I have edited out the comments about Blu-Ray. (They were somewhat vague anyway, since I did not then have access to the details.) The bottom line is that the Blu-Ray DRM system is set up to not work unless you pay — not only for the disc but also for the player ( https://www.howtogeek.com/240487/how-to-play-dvds-and-blu-rays-on-linux/ ). I installed a free version of one commercial player (software), and it worked perfectly, albeit with a large text banner in the centre of the screen all the time.
______
— If you have just joined us, you can jump straight to Updates — the last material. […Although there are questions in the thus-skipped section.] —
************ BTW: I ask a lot of incidental questions herein, but the core question is that of what is wrong with my machine and how to fix it.
I have not taken notes detailed enough to be adequate for it, but here is the situation from memory. Sorry if this seems excessively detailed, but it is that kind of problem.
This should be irrelevant (I imagine), but:
Motherboard — Gigabyte GA-Z170X-UD5 (*not* TH)
CPU — i5 6600K
RAM — as per text; currently G.Skill F4-2400C15D-16GVR, running without XMP (or whatever it is) at 2133 [thanks UtterDisbelief].
Edit<
Storage — an excessive collection of 3TB HDDs. My boot volume is the same for Windows, Linux and MacOS — all manually assembled. [I should be able to answer basic questions about configuring Windows for multiple boots; the secret is: duplicate, not create new entries.] That is, I have all my bootable partitions on the same drive — a 3TB Western Digital Blue. (I somehow ended up with a spare of the same type; at one point recently I was trying to clone my boot drive onto that.)
>
Let me note at the outset that the machine is perfectly happy at the moment in Windows or Linux; it is only in the Hackintosh that it freezes. Also, it still freezes under Safe Boot.
<This part should (*shoouuuld!*) be irrelevant. (I include it for completeness’s sake.)>
I originally bought some fast, large RAM — 32GB@3200. [This was a snap decision that I regret.] (It was F4-3200C16D-32GVK, but that does not matter). The point here is that it was not listed against my motherboard — neither on the site of the motherboard manufacturer nor of the RAM manufacturer (being Gigabyte and G.Skill, respectively)… but I did find someone on the internet who had that combination and reported that it worked fine.
As I was setting up my Hackintosh, it did not faze me that it crashed here and there. Finally, however, it became apparent that it was crashing — freezing, particularly — once every one to two weeks, arbitrarily — including when it was ostensibly idle.
I found out (rightly or wrongly) that this is a RAM issue. I bought TechTool Pro and ran it, and it reported a stuck address. (Note that this runs under the Hackintosh OS. That is to say that, at this point, I had not actually put in the RAM settings in config.plist; I would have gotten around to that much earlier under different contingencies. Of course, I should have realised that running the check under MacOS was therefore not a good idea. I did actually think of this, but I thought that the program would talk directly to the RAM. (Maybe it does!) )
I took it back to TechBuy, and they duly returned it to G.Skill. I was told that it would take about three weeks, so (for reasons I can not recall offhand) I felt it necessary to buy some RAM for the interim. [Edit: … and I installed this.] Actually, I had the replacement RAM back in about three days.
I found out later that TechBuy had written, “failed memtest”, on the paperwork; I now infer that G.Skill took my word for it that the RAM was indeed faulty; I was actually hoping that they would check it themselves. My guess, now, is that it was fine — just incompatible.
It turns out that you can download the excellent “memtest” for free, and it boots itself and everything. All my later checks were with this.
I did not bother to actually replace it [i.e. install the replacement instance of the original RAM] in the machine, and ultimately it was about six months before I did this. (Everything has been a nightmare for this whole period.)
The point is that I had the substitute RAM (F4-3200C16D-16GVKB) installed in the machine for six months, and it was fine as I recall (or maybe it only froze every four weeks or so)… and this substitute RAM is also *not* listed as tested to be compatible — a stupid and embarrassing mistake on my part. [The various RAM is “original” and “replacement” for the 32GB@3200, and “substitute” for the interim substitute that ran for six months (all G.Skill).] [By the way… if anyone knows why RAM manufacturers have to test their RAM against individual motherboards to see if it actually works, I would be very curious to hear how that makes sense. (Do they genuinely not know how?!!) ]
Recently, then, I have installed the replacement RAM in the machine, and it was not okay — I ran memtest, and (from memory) it reported that it was fine on the first run, but on the second it reported a collection of errors at the top end, all on the same bit in a 32-bit word (or whatever it is). I have been told that that is characteristic of incompatibility. I think the machine was crashing about once a week at that stage.
I then tried the substitute RAM; I can not recall the details, but I would have run memtest on it, and the outcome was that I raced out and bought some (!!certified!!) RAM. (If memory serves, it was crashing, and I ran memtest and it reported no errors, and finally I ran it again and it did report errors.)
</This part should be irrelevant.>
The outcome of the foregoing is that I now have compatible RAM in my machine, and I have run memtest on it twice to make sure.
So…
This is where I would have liked to have comprehensive notes. Around this time [i.e. during the ultimate sorting out of the foregoing RAM issues ((or believed I had?)) ], the freezing went from under a week to under 24 hours. In hindsight, something must have changed in the software side, during this period; I did not realise at the time that this represented a systematic change. [See also “Notes”, below.]
I am thinking that what has happened is that I have fixed a freeze caused by incompatible RAM, and coincidentally developed a software problem.
There are only three candidates that I am aware of — my best guess is the third. This is *not* to say that I am in the slightest ruling out some other cause.
One is that I tried hard at some point to get software that would play Blu-Rays. (I had found out previously that this is not a trivial exercise.) I have installed a variety of players to try them. (I have Kaspersky anti-virus software.) I am sure that I would not do this when my computer was crashing every 24 hours. Conversely, I have a vague memory that it was a couple of weeks or more earlier, that I was fooling around with Blu-Ray players — the point being that that would rule that out as a cause. Still a possibility.
The second candidate is the audio. Again as described below, I came to suspect that this might be the source of the problem, and had trouble trying to fix it. At this point(as I write), the AppleHDA.kext is in the Trash, and everything else should be normal for a Voodoo installation. [My understanding is that Voodoo has a disabler for the AppleHDA, which should imply that Voodoo would be working.]
The third candidate is double-barrelled, as I understand it. I found out — through doing a safe boot (which checks the hard discs) — that I had corruption on my hard disc drive. Initially, it did its thing fixing it, and then reported that it could not fix it, and turned off the machine. When I turned it on and did another safe boot, it tried and said that it had fixed it. I know enough about these things to know that it is not fixed — that the first run made some changes that locked in an error.
The point is that it was already in (what I am taking to be) the new software crash era, before this “fix”ing, and that after this it was much worse — it went from crashing within 24 hours to crashing during the startup phase — the part where it has the desktop showing, and is pretending that it is up and running, but actually takes a few minutes to finish starting everything up. It does not always crash during startup; it ranges between that and five or ten minutes later. (The best is about 5 hours, but that is a massive outlier.)
This brings us to the recent past.
I came to suspect the audio — Voodoo 2.8.8 — on the basis that… (since I reinstalled it ages ago) it always comes up with a message that it could not load the Voodoo kext (plus a few coincidences). There were 2 instances of the kext — one in S/L/E* and one in L/E. From memory, offhand… I deleted the one in S/L/E, and later/soon came back and moved the other one from L/E to S/L/E. (I remember being surprised that the sound suddenly came on while the machine was running; that would be this; (at least on my machine) it has to be in S/L/E.) Early on in this episode, I reinstalled El Capitan and the Multibeast set. Particularly, I tried installing the Series 100/ALC1150 version; it does not work on my machine, but I thought that perhaps it [the machine] would not crash continually.
[ * <hard disc>/System/Library/Extensions ]
Finally, I trashed AppleHDA.kext from L/E. (In hindsight, this was very foolish, but at the time I was thinking that maybe it was the AppleHDA.kext that was causing the problem. [As well as I could tell, the backup and the original were exactly the same — which I thought was rather strange; there is no point carefully replacing the original if the substitute is no different (not to mention having a substitute in the first place).] )
I think my problems have worsened significantly from that time.
My project at the moment is to do a low-level backup of the machine as it is, and swap in a copy of the partition from a while ago (plus all the attendant work stuffing around with my work that is thereby deleted), but that endeavour, too, seems to be cursed.
In summary, then, my Hackintosh freezes under safe boot, after reinstalling the OS and the MultiBeast stuff (but the hardware is fine under another OS).
Any help would be deeply appreciated. (It would be *far* less of a hassle to fix one software problem on the current setup, than to roll back to a backup from several months ago and then have to transfer back in all my mail and documents, plus all the work I have done on the machine in the interim.)
_________
[Context: I [will] have backups of the partition as it is now, and as it was several months ago.] _ One concrete question… . Assuming that it is something related to either • audio drivers or • software installed by a Blu-Ray player… given that I have a Clonezilla backup of the partition from before the (new) freezing started, and assuming that I can mount this and (Linux forbid!) actually copy the files… is there some group of files that I could copy, that should get rid of the problem, such as to save me reinstalling/transferring all my email and browser history and bookmarks and documents and what-have-you, please? (This would save me a *lot* of hassle and time!) For instance, if I were to copy all and only • System(/Library) and • Library, then that would delete all my mail, bookmarks and so on [or possibly some of this is stored in my User folder?], but if I were to copy only (say) S/L/E… would that do it? [p.s. I do not understand why copying the AppleHDA.kext back does not fix it; it there any form of copy, short of a partition replacement, that will work?]
Speaking of which… I have posted a question about this online, but I might as well ask it here as well. Linux flatly refuses to have anything to do with files such as those copied from a Macintosh into a backup, on the grounds that the permissions prohibit it. (The source has the wrong owner, and the target is read-only.) Ignoring the issue of having to learn command line flags for every little thing, and wondering if one has made some terminal mistake… the “solution” seems to be to actually change the ownership and permissions of the files. (a) If I do this to system files, would a permissions repair put it all right again? Is there (say) a hacked version of Linux that is not obsessed with checking permissions? (Can I set up a user account called “root”, and everything will just work?) On the Windows front, the problem is that the geniuses at Microsoft put a limit of (about) 256 characters on path designations. Is there a hack out there to fix that little oversight? [Later: apparently not.] (Apparently there is in Windows 10, but I am using 7. I am just trying to see if MacDrive will prevent this. I have Paragon’s HFS+ thinggy (which does not).)
Incidentally. Is it theoretically possible to put together a (presumably large (at the time of writing)) USB stick with a bootable HackOS (including the Finder) plus the ability to (a) copy Dd and Partclone *partitions* and (b) the ability to copy *files* from said? If so, how far is this from immediately practicable, please? (I have just been searching; I found this https://www.insanelymac.com/forum/t...ow-to-make-a-bootable-leopard-on-a-usb-stick/ (from 2008), and then stopped.)
If so, can it be given the ability to read dd and partclone containers — this being the format of my backups (done with/by Clonezilla)?
Notes
As far as I can tell, there is nothing in the Console logs, ever. (There are excessive wake-ups going on, but I take it that is not about sleep (as in low power), and that it is pretty much normal.)
I should note that I have a Skylake processor, and I am only on El Capitan, and my machine setting is “17,1”, so my operating system does not know about the processor. I tried once telling it that it was “14.2” (without making any other modifications), but it still crashed a moment after booting.
I have tried upgrading to Sierra, but, in most of my attempts, the machine has crashed before it finishes the 1-minute setup, and in the one case in which I got to the reboot point, the {continue with installation} boot option was not there. (I tried booting normally, but that did not work.)
I have been trying to get Blu-Ray movies working on the machine, and I have installed all the usual suspects, including one that requires Java Runtime (SE) 6. Perhaps unwisely… I found a terminal script, obviously written by someone who knew what they were doing, that purported to get VLC working properly… that (among other things) changed the permissions of some pertinent files.
Informing the third candidate, above: I have been amalgamating instances of my stuff (applications, utilities and documents) from about four sources — i.e. various past machines — preparatorily to copying elements (as is indicated) of this amalgamation onto the machine in question. The point is that I may have inadvertently copied something corrupt onto my gleaming new machine — for what it is worth.
Updates
20180501
Trying to boot and shutdown cleanly in MacOS, so that Clonezilla will not balk at copying my partition. Instead, I got many, many instances (most the same entry, oddly enough), of:
hfs_swap_BTNode: record #38 invalid offset (0x0000)
hfs: node=16199 fileID=4 volume={My boot drive} device=/dev/disk5s12
hfs_mark_inconsistent: Runtime corruption detected on {my boot drive}, fsck will be forced on next mount.
Also:
<Emergency> Boot task failed: fsck-safe
By the way: I ended up with 2 instances of a HD Recovery partition. That is not going to cause any problems, is it?
I guess my next move is to reinstall El Capitan again, and try again to get a clean (boot and) shutdown.
That is… it is looking more and more as though there is something actively screwing up my disc (that is, this volume), and that the best alternative is to nuke it with a previous image and go through all the hassle of updating that.
I do have some corrupt files copied from failing previous machines, but they are on different volumes. I have been copying files onto the new machine, but there should not be any corrupt files among those (surely!?). …And even if there were, they would not actively damage the disc after it was repaired.
Thank you all for taking the trouble to read this!
… I am developing a strong-ish suspicion that the Linux endeavours to repair my hard drive (preparatory to Clonezilla backing it up) are what is causing my immediate problem (or are part of the problem or whatever). Last time around, when I was trying to do a backup with Clonezilla, I told it to do repairs as necessary [noting that there *were* errors], and I remember that it passed the first three layers without comment… and when I next looked at it with Disc Utility, *that* failed absolutely to make any sense out of it, and gave up. (I would have thought that if HFSProgs was bad, I would have stumbled across mention of it several times in the past.)
… Or maybe it does not understand Journalling, and it is all my fault for trying to use it to do repairs. Trying Dd.
<Edit/Note:
In my desperation, I accidentally allowed a Linux program to do a repair of an HFS+ partition with Journalling still on, thereby making it much worse… noting that, now, I would not choose to do this at all (even with Journalling turned off). I am almost sure that my hard disc was already corrupt, at this point (which would be why the Linux program wanted to fix it). It would seem that Journalled HFS+ is *not* merely HFS+, plus historical notes — such that one could work with it as though the notes were not there.>
[For what it is worth, I note in my defence that GParted explicitly notes that there are unapplied journal entries, and offers to Fix this (and apparently fails to without bothering to report any errors); this seems to me to be inconsistent with it not supporting journalling.]
/ Edit/Note>
20180503
Okay; this isn’t funny any more (not that it was before!). I have just cloned my hard drive onto another one (so that I have some hope of recovering my files from the broken file-system), and replaced the contents of the HackOS partition with a partition backup from a few months ago. [For avid readers: that is from the period when I was still thinking (correctly, I believe (but not absolutely) ) that my problems were caused by incompatible RAM.]
It booted fine… Disk Utility said the partition was fine… and it still froze after 1hr04min. The RAM is fine. (I have checked it twice.) The only thing apparently wrong is that it said that it could not run the VoodooHDA.kext — the reason for that being that I have ended up with 2 copies — one in S/L/E and one in L/E. Arguably, if it refused to load it, it can not be causing any problems.
! If I failed to re-install AppleHDA.kext (or whatever the Apple sound thinggy is), but instead merely copied the file back in (not realising that that was not adequate), would that cause this problem, please, anyone? (Just out of curiosity: if that obtains, can I get the machine running by deleting the thing entirely?) I suppose that would explain why the installed copy was exactly the same as the backup copy — that is, that I had previously copied it back after trying ALC1150 and then changing to Voodoo. If I install ALC1150, would that disable AppleHDA and fix everything — if this is indeed the problem, please?
Again: the machine is perfectly happy at the moment in Windows or Linux; it is only in the Hackintosh that it freezes. Also, it still freezes under Safe Boot.
Any help greatly appreciated; thank you in advance.
20180512
(After a struggle) I managed to get Sierra running; it has run without freezing for over 24 hours, at least twice. {Emoticon for laughing hysterically then fainting.} (I get a sense of glee and deep satisfaction from moving the mouse pointer around. Sad!)
The really intriguing thing is that it freezes (as above) while the Sierra installer is running (both while booted from the USB (e.g. while installing) and while booted from and still setting up the new OS). Apparently there is something about completing a (Sierra) installation (and rebooting) that takes it from freezing to not. [Just repeating: there is never anything in the logs (unless I am not reading them correctly).]
It was indicated, ultimately… that there was some corruption in my El Capitan installation… but that does not work for booting from a Sierra installation USB. (Supporting the foregoing tenet is the fact that it was (mostly?) okay for months, several months ago.) Unless this happens to everyone, the indicated conclusion is that there is something in my hardware that is fine under Linux, Windows 7 and Sierra, but not under El Capitan nor the Sierra boot environment. … Something entirely unique to my (particular instance of my) hardware. Even if I imagine that (for instance) my motherboard is faulty, it is hard to make sense of it all — not that I am an expert on hardware fault symptoms. It is perhaps more likely that there is some anomaly in the Clover settings on the installation USB (plus coincidental corruption of my El Capitan install)… but then there remains the question of why this affects only my machine (GA-Z170X-UD5, i5 6600K).
? Did anyone else with this (or any other) motherboard get arbitrary freezing during installation?
I shall probably never know what it was. (It would be interesting to do a clean installation of El Capitan, but that is certainly not a priority right now. I expect it would work fine… .)
Last edited:
