DaveOrAstroDave Posted October 3, 2018 Share Posted October 3, 2018 Hi, I installed Mavericks on a Haswell system (i7-4770k on a Gigabyte Z87-MX-D3H & GTX 770 GPU) and have upgraded it successfully through every version of macOS since then, up to High Sierra. I've kept Clover up to date, and thought I'd give Mojave a try. The install went smoothly, but now it takes FOREVER to boot, and I get a beachball every few seconds. I checked the syslog, and aside from a ton of "class is implemented twice" messages, the most suspicious thing is that configd crashes with a sigsegv every 10 seconds. In the stack trace, the crashing thread appears to be doing something related to setting the IP address. I've tried removing my wifi card (one of those Apple branded Broadcom devices in a pcix adapter) and disabling my on-board NIC, but configd just keeps on crashing. If I manage to get terminal open, I can stop configd w/ launchctl, and then all of a sudden my system is super responsive again (with no networking) until the next reboot. I've looked for suspicious/old kexts, launchdaemons/agents, etc, and every lead is a dead end. Does anyone have any ideas what else I can check? Sorry if this is a little thin on details, I'm not home right now. I'm happy to provide any useful debugging info when I can. I really appreciate any insight anyone can provide; I'm banging my head against my desk. Link to comment Share on other sites More sharing options...
Quirkz Posted October 6, 2018 Share Posted October 6, 2018 interesting, I've seen exactly the same thing - And found your post by searching for 'confided crash mojave'. What (may) make you feel better is that I'm having it on an actual iMac and a macbook pro - Not a hackintosh. There are threads in the apple forums which detail the same symptoms on a mojave update, but little to resolve it. Touch Bar not working, networking failing, bluetooth devices disconnecting, slow computer, etc. It happened to me on *both* my desktop and laptop that I upgraded. Ended up rolling back to High Sierra. This makes me think it's environmental, and something I set up or whatever on those machines triggered it. So, if you don't mind, maybe we can compare notes and try find some possible commonality: * Any dev tools installed? I've got homebew and some databases, and Xcode. * Any VM's? I've got docker running with it's built in virtualisation * How old is this install? Mine have been transferred from Os upgrade to upgrade, and from machine to machine - And the installation is 'older' than the machines themselves. Any other ideas that might trigger configd to constantly crash, or do you think that configd crashing is just a symptom, rather than the cause? Link to comment Share on other sites More sharing options...
Quirkz Posted October 6, 2018 Share Posted October 6, 2018 Link to apple threads discussing the issue - https://discussions.apple.com/thread/8552955 Link to comment Share on other sites More sharing options...
Quirkz Posted October 6, 2018 Share Posted October 6, 2018 Question: Do you have a /etc/sysctl.conf file? If so, what are it's contents, and what is the date on it? Link to comment Share on other sites More sharing options...
DaveOrAstroDave Posted October 7, 2018 Author Share Posted October 7, 2018 Wow, it feels good to know I'm not alone. I haven't had much time to put into this yet, but I'm happy to answer your questions: 1) Yes. I've also got homebrew installed, along with xcode, android studio, Unity3D, and maybe some more stuff. 2) Yes, Parallels desktop. But I already uninstalled it to see if that would help. It didnt. 3) It's an old install. Not migrated from a different computer, but old. It was a fresh install on the same hardware back when Mavericks came out. 4) I'm not sure yet, but I tried starting configd manually with various testing/debugging options, and I've found that if I load individual bundles with it it won't crash, unless I load the KernelEventMonitor.bundle. Unfortunately configd seems to be useless without it. I think it might have to do with another error I'm seeing in the log regarding a libdispatch bug. Of course now that I say that, I cant find it in the log anymore... I do have an /etc/sysctl.conf: # (comments) # START kern.ipc.maxsockbuf=512000 net.inet.tcp.sendspace=256000 net.inet.tcp.recvspace=256000 net.inet.tcp.delayed_ack=0 #END the date on it is Apr 11 2009. Suspiciously old, eh? Link to comment Share on other sites More sharing options...
DaveOrAstroDave Posted October 7, 2018 Author Share Posted October 7, 2018 Solved! sudo rm /etc/sysctl.conf sudo rm /etc/appletalk.cfg sudo rm /etc/appletalk.nvram.en1 (not sure which one did it) basically i compared against a working laptop install and deleted stuff that wasn't present on the laptop. I still need to put various hardware back in (gpu, usb/fw, wifi, enable nic) but configd is now running without a crash and the system starts up crazy fast. I'll report back once I get the rest of the hardware in. THANK YOU!!! Link to comment Share on other sites More sharing options...
DaveOrAstroDave Posted October 8, 2018 Author Share Posted October 8, 2018 Confirmed, my machine is 100% back to normal with no apparent ill effects. Thanks again for the lead! Link to comment Share on other sites More sharing options...
Quirkz Posted October 8, 2018 Share Posted October 8, 2018 yeap, glad the hunch paid off. I'd noticed that my two machines that failed had a sysctl file, while new machines no longer have the sysctl file. Removing the sysctl file was the one that did it - That's all I removed on my machine and now it's also resolved. Link to comment Share on other sites More sharing options...
ice_pdb Posted November 15, 2018 Share Posted November 15, 2018 @DaveOrAstroDave Hi! would you comment how did you update?... I'm running a similar generation PC and I want to know about update steps. Thank you very much. Link to comment Share on other sites More sharing options...
Recommended Posts