CB's instability issues
Okay so. A few months ago, my PC started crashing intermittently. At first it was only once a week or so, so I didn't think much of it, but it kept escalating until it was every day, then multiple times per day, and I am at the stage where it feels like I might have to just throw the whole thing out the window and start over, but I hope there are some more intermediate step I can do.
So, first thing, I thought it must be heat, maybe something about the new flat was making it run hotter? But no. I monitored and logged the temps within the case and with that core temp monitor for weeks, and nothing was running outside of normal operating parameters.
So, I thought it might be memory, so I ran Memtest86 (CLASSIC STUFF HERE MAN) But it ran all night and came back clean, so not the RAM.
So, I reinstalled Windows. At first it seemed to help, it went like a whole week without crashing, but then they started getting more and more frequent again, and now it's back to 2-3 times/day.
This leads me to believe that something about some combination of apps or app settings is the culprit because it didn't start crashing as much again until after I was really back up and running all the same stuff I was running before with all the background apps and everything.
SO: First, is there any way to figure out what is going on when it crashes? I'm pretty sure there is not, but if someone has some esoteric way to figure out which bit of software is causing my PC to hard-crash on the reg, that would be super.
Otherwise, it seems like my next step is to start swapping out bits of hardware, but I don't really have a bunch of hardware laying around these days, so I would prefer if there were other ways to test the hardware (like Memtest does for RAM) to see if it's the processor or the mobo or the GPU or what dumb bit of tech in my box is doing this to me.
Any help is appreciated. ( I can't live like this anymore)
Gigabyte Z77X-UD5H mobo
Intel Core i7-2700K CPU @ 3.50GHz
GeForce GTX 1060 3GB
16GB RAM (I know 8 of it is Crucial DDR3L-1600, but I don't remember the other 8)
Do you have a spare PSU? If so I'd try that to isolate the PSU.
You can also pull the GPU and run from the integrated graphics to isolate GPU.
I'd also recommend running one stick of RAM at a time to see if the crashes happen with one stick but not the other.
Parenthetical aside: Good thing you have those great windows that open all the way out for when you chuck it.
Download this ISO and make either a CD or bootable USB from it (Rufus is one tool for windows that does this)
When you boot, it is a relatively straightforward menu of options to test different components. Stressing CPU and GPU and the system hard-crashing could definitely indicate an issue with power delivery. If testing individually doesn't yield anything, you can do testing in Windows with both a CPU and GPU stress test at the same time using a variety of tools, but Prime95 or OCCT (https://www.ocbase.com/#hero) are good for CPU; For GPU, try Furmark (https://geeks3d.com/furmark/). OCCT also has a GPU test, but I don't think you can run both CPU and GPU simultaneously (for max power draw).
In terms of figuring out what is going on when it crashes, have you reviewed the System and Application logs to see if there are any consistently reported errors being timestamped as occurring just prior to the crashes? If so, this would help narrow down if one or more apps or services is involved in the issue.
Memtest86 is no longer helpful. You need MemTest64 for Windows to get an effective memory test these days.
Given the highly advanced age of your motherboard, you may also want to do a visual inspection for blown or leaking capacitors.
I had a problem like this a few years ago and it was caused by the virtual memory on an aging, close to the end of its life hard drive.
I switched the system virtual memory to a different drive and the crashes stopped. The whole OS got switched to a new drive after that
@GHoosdum : I don't have a spare PSU, BUT your comment made me think about power delivery, and it made me wonder: Is it possible that the shitty wiring in my new flat is to blame? (GH already heard about it, but for others: My flat only has two circuits for all the non-specialized outlets, and each caps out at 3k before it trips, so my PC is on the same circuit with like, the fridge (and everything else in the kitchen other than the oven) and the hair-dryers in the bathroom, and the a bunch of other junk in the flat.) It occasionally causes problems when the breaker flips, but could that wiring also cause the PC to be under-powered and crash? I need to get a UPS anyway because I don't like losing power to my PC when the family decides to blow dry hair and make a smoothy at the same time, but might that also be a solution for crashes?
Also: Isolating GPU added to the troubleshooting to-do list.
@Tushon : Currently downloading. I'll try it presently.
@Colgere : I didn't know these logs were a thing. I'm looking at them now, and don't see any particular pattern over the last few crashes, but I'll be keeping an eye on the logs after future crashes to see if I start to spot anything.
@Thrax This was an easy one to just do when I went to bed last night, so I did it. 8 hours: zero errors, just like MemTest86. I'm glad to know that there is a new testing app though. I didn't realize that 86 was too old now.
@MrTRiot That's a good idea, and easy to try. I just switched my pagefile from my older SSD to my newer SSD. Even if it doesn't stop the crashes, it's probably better that way.
A UPS is a good investment for any computer (that isn't a laptop), so I recommend that regardless of your troubleshooting efforts. Power delivery fluctuations at load could certainly be related.
Changing the Pagefile to a newer drive didn't help. The system has crashed hard 5 times since making the change 15 hours ago.
Tried the Ultimate boot tools, and I couldn't get them to work. I know how to make a bootable USB drive, I did it for Memtest, but These instructions are either poorly written or written for a different audience because I think I'm having trouble parsing them and getting it to do what it's supposed to do. I'l the try the windows based tools next.
I've been checking the logs, and four of the five recent crashes had a fatal application error recorded ten seconds before the crash. One was in a file called Lcore.exe, which I can't find on my system to know what it's associated with. The googs seems to show that it's an OS file, so maybe it's hiding from search (though I have all 'hide things from the user' options turned off. The other three errors were with an exe that is associated with my always on back-up service that is supposed to be keeping a copy of all of my files duplicated on their servers. It's one of the few apps that is always running whenever my PC is on. I can't tell if it's causing the crashes or if the crashes are causing it to fail in some way, but I'm going to try shutting it down for a bit.
re: UBCD, if you're using rufus to make the USB, it's a pretty straightforward select ISO -> check bootable and make sure your target system is correct (it probably is, since BIOS has been around forever, but you might need to tweak partition scheme). Otherwise, please post what you're trying and what isn't working. I've had some trouble with other USB creation tools in the past, so I would highly recommend Rufus.
You may want to run
sfc /scannowfrom an elevated command prompt to check on your core windows files (but this isn't commonly the issue).
@CB rufus needs to be set to MBR for partition scheme instead pf opensomethi ng for any windows based isos. The instructions for it are very poorly written..
Also I found this link relating to Lcore.exe (https://www.file.net/process/lcore.exe.html) which might give you more of an idea of where to go to next...
Have you run malware bytes yet by chance? Might be worth a shot depending on how Lcore.exe is behaving...
Ah, so LCore is part of the Logitech package, one of the other apps that is basically always on, since it is what handles the customization of my mouse and keyboard. Weird that it didn't show up when I searched my whole PC for the filename.
I couldn't grok the instructions on the UBCD site, so I just made a bootable USBdrive (with rufus) then copied the files from the ISO onto it, and restarted my PC. I told BIOS to boot from the USB drive, then it just booted into Win10 normally. :/
I don't see an option in rufus to specify an ISO. Edit: N/M found it. I'll try that.
Still can't get UBCD to work, but I ran OCCT for five hours with no errors.
It's very strange that Lcore didnt show up when you looked for it. Strange enough that I would run a complete malware bytes scan to rule out something nasty like a rootkit.
If that doesnt show up anything it would be a good idea to start taking a deep dive on potential electrical issues (frayed wiring, loose fan connections, inconsistent wall power draw, ect....)
You could also attempt to look at your ram usage right before a crash and see if you have a pretty epic memory leak.
The PSU arrived yesterday. I opened up the box today to find that the plastic case had been cracked, and was buckled outward. Immediately packed that back up for return. The brand had good reviews, but considering ordering from somewhere else after that.
@MrTRiot That's a good idea. I'll DL malwarebytes presently.
Memory leak was one of the first thing I checked along with the temps at the beginning. I just failed to mention it in the OP. My RAM usage while not under significant load is never over 40%, and that's when the crashes typically happen (oddly it almost never crashes while I'm watching a film or playing a game. It's usually while I'm doing some low-load task or when I'm not at the PC at all). HOWEVER, that was only a visual spot-check. I was keeping the resources monitor up on my second screen so that when it crashed it would freeze-frame a record of what the system load looked like, including RAM take-up. I'm not sure if that's a super accurate way to check for a memory leak.
Malwarebytes only flagged a few Chome extensions that I installed on purpose.
I hope you meant Chrome, and for first scan you need to upload latest defs.
Um, you might have RAM with a flaw in a stick. memtest64 tests for that.
Finally got the new UPS set up. We'll see if it helps.
I hope it is now stable.
It's not. I'm currently going back and forth with BackBlaze support staff because turning off BackBlaze makes the system more stable, and I'm giving them a chance to help me find a solution before I change cloud back-up providers over it.
Does your data get encrypted before it gets sent to backblaze? I've fixed a few computers with end point encryption and its causes a whole lot of problems.
Yeah, it's encrypted on my end first.
I no longer think it's BackBlaze. I had noticed more stability, but I think it was just random. It basically has an uptime that randomly falls between 15 min and 72 hours, so it's really hard to tell when something is working better. Sometimes I'll do something, and it'll be stable for three days, and I'll be like "FINALLY FIXED" but then it'll crash again in the exact same way, and then do it three more times that day, and well...
Frankly I'm just done with it. The oldest parts of this Frankenstien's machine are nearly 15 years old. I think it might be time to just start fresh with an all new build.
I'll start a new thread for that.
Intel has some good processor utilities and some of them test other sub-systems. They are very good at finding issues. My wife's machine would pass all but one of the modules, and it pointed to internal power issues (time for a new machine for her).
App related stability issues are a nightmare.
Note: I was glad to see that your post started with "My PC..."
Form the title of the thread I was a bit concerned about you .........