09-28-2019 09:27 AM
Hi, my ThinkPad P1 has been randomly freezing for some time now. There seems to be no apparent pattern to when the freezes occur, as I've tried to coorelate the freezes with graphics adapter (hybrid or discrete mode) in use, heavy or light CPU load, heavy or light disk activity, Internet browsing, and so on. When the freeze happens, it's unrecoverable; only way to unfreeze is long press of power button to restart.
Here are the various steps I've taken (unsuccessfully) to address this:
- I've had two service requests (I purchased the warranty plan) to replace the main board
- I've tried both hybrid and discrete graphics modes in BIOS
- I've upgraded the BIOS to latest version
- I've toggled various BIOS settings with CPU and Memory (disable vt-x/vt-d, disable memory protection)
- I've booted with a Ubuntu Live USB image (freeze still occurred)
I've had no success with any steps above, in particular two different replacement mother boards.
At this point I'm requesting a new machine, as this problem has become way too disruptive to manage (unsaved work is lost, and no telling when the freeze will come). I asked the tech at the 2nd service appointment what next steps would be if the latest main board replacement was ineffective, and he seemed to agree that a replacement laptop would be a suitable follow up to the failed main board fixes.
1. Anyone else experienced this behavior with a ThinkPad P1 of late '18/early '19 vintage
2. Anyone had success with a machine replacement through the warranty program?
09-30-2019 06:38 AM
10-01-2019 02:52 PM
Sounds like a memory issue. How much RAM in the system? Since you had the system board replaced, the memory is suspect IMO.
10-02-2019 08:27 AM
Hi @harrisb, thanks for your reply.
That is a good thought. I believe I had at one point ran memtest86 to confirm the issue wasn't memory related, but will try again. FWIW the system RAM is 32Gb as 2xDIMMs if I recall correctly.
In either case, RAM or not RAM related, the fact that support sent me a new NVMe doesn't seem pertinent at all to this issue, when you consider I still saw the freezes running off a USB ISO. At this point, having had 2 new mainboards installed, the issue IMO is either memory related, CPU related (keep in mind that the service tech did actually re-insert my CPU into the new mainboards; I saw him reapply the thermal paste), or perhaps even BIOS. It is not anything to do with the SSD/NVMe. I pondered simply not opening the NVMe they sent me and just sending it right back, but I'm concerned they are going to stonewall escalation of this case unless I follow their steps (as illogical as they may seem).
Speaking of escalation; does anyone have any ideas on how to expedite a warranty claim to just get a new system? My patience with the whole situation is wearing thin, after 2 on-site service replacements and now a (seemingly hoplessly flawed) NVMe self-service delivery.
10-02-2019 08:46 AM
The CPU is soldered to the system board. I don't believe that the tech would have removed it from the original board. I believe that you have a memory issue. Try running with just one ram module and see if the freeze happens. If it does, remove that module and put in the other one. If no freeze, then the first ram module is bad.
10-03-2019 04:52 AM
Yes @harrisb, you may be correct. I did observe the repair tech applying thermal paste, but that would occur in either scenario, soldered or pin-connected, between the CPU and heatsink.
In any case, I've put in the new NVMe, and ran some HDD tests and `dd` to zero out the new drive. So far no issues (as expected). I'm going to run memtest86+ on it next, and depending on the output of that, will put an OS on the new drive.
10-03-2019 02:46 PM - edited 10-03-2019 02:58 PM
WOW, so after executing memtest86x+ via an Ubuntu Live ISO, I observed many hundreds of memory errors were reported. The test run lasted many hours, but the program appeared to actually freeze (the memtest86+ app discontinued counting elapsed time or increasing its test % completion) at about 67% way through. I ended up powering off the machine 4-5 hours later when it became clear no additional progress was being made.
I snapped a photo andattached below, to see if anyone has any specific comments about the (partial) results. I looked to see if there was any way to figure out if either of the individual DIMMs was faulty (while the other was OK) but based on the progression of the tests, it appeared to me that errors were reported all the way up the address range, 0 - 32Gb (I had seen a range test of 21Gb-26Gb in an earlier test), which implies to me that both sticks are compromised.
Nevertheless I will bring this up in my report ticket and ask for RAM replacements as a next step.
10-03-2019 03:18 PM
One DIMM at a time as I've previously suggested.
10-08-2019 03:12 PM
To update this thread, I have gotten in touch with Lenovo support concerning this issue. The next step in the support process is apparently sending my laptop to depot repair for evaluation. Supposedly if they are unable to diagnose the problem and repair it after this service request, since Lenovo has 3x service incidents (they don't count the support request where they shipped me an NVMe?), they'll subsequently replace the laptop with a new one.
I'm dissatisfied that they are not counting the shipment of the NVMe as a service incident. I never requested them to ship a new NVMe to me in the first place; they took this step unilaterally without any discussion from my end. A cursory review of the case (where I described that the issue occurred even booting the machine from a Linux USB thumb drive) would indicate the problem had nothing to do with the main OS drive.
In any event, I have no choice I guess to follow the process and ship out my machine for repairs. I've been without my laptop for nearly 2 weeks, and the expected turnaround with depot repair is another 1 week (if all goes to plan). Hopefully this is the final service request, as my patience is wearing thin.
10-24-2019 04:11 PM
Not sure what happened, to my last post but will retype it.
My depot repairs did NOT address the freeze issue, unfortunately. A manifest returned with the laptop says depot repairs included a new main board + CPU, and new RAM modules. I did manage to verify the RAM modules using the memory diagnostic tool built into the Lenovo BIOS, so I believe those were replaced with working units.
However, the machine actually locked up during an attempt to save the BIOS settings yesterday with no OS at all running. I did go ahead anyway and install two different Linux distros to the machine (Ubuntu and Fedora). Both of the USB installers for these distros locked up before completing the install. After several attempts, the Fedora install worked. However, the lockup is repeatable now; it occurs everytime I login and reach the desktop windows.
I continue to work with support and am requesting an entirely new machine. I will update this thread with news as the process unfolds.
Appreciate and tips/experiences from those who have had full Lenovo laptop replacements before.