You are not logged in.
Pages: 1
Hi, I have a computer which I mostly use for Deep Learning, and since the last 3-4 weeks, it is rebooting randomly. I connect to this computer throush SSH and I keep my sessions active using Tmux.
You can see here my reboot logs:
last reboot | less
reboot system boot 5.9.8-arch1-1 Wed Nov 18 05:47 still running
reboot system boot 5.9.8-arch1-1 Tue Nov 17 09:22 still running
reboot system boot 5.9.8-arch1-1 Tue Nov 17 04:53 - 09:22 (04:28)
reboot system boot 5.9.8-arch1-1 Sat Nov 14 16:52 - 09:22 (2+16:29)
reboot system boot 5.9.6-arch1-1 Mon Nov 9 16:08 - 16:47 (5+00:38)
reboot system boot 5.9.6-arch1-1 Mon Nov 9 12:09 - 16:07 (03:58)
reboot system boot 5.9.6-arch1-1 Mon Nov 9 10:35 - 16:07 (05:32)
reboot system boot 5.8.14-arch1-1 Mon Nov 9 09:57 - 10:22 (00:25)
reboot system boot 5.8.14-arch1-1 Mon Nov 9 01:00 - 09:57 (08:56)
reboot system boot 5.8.14-arch1-1 Thu Oct 22 09:09 - 09:57 (18+00:47)You can see that it rebooted today at 5.47am. I don't understand why it says that the previous boot is "still running".
You can see here my journals before the boot: https://termbin.com/qcug
And my journal after the boot (still the active session): https://termbin.com/jzrbb
As I said, I mostly use this computer to do Deep Learning. The strange thing is that previously, when the GPU failed, the system would not reboot, the learning would fail and the gpu would be off, but the system would not reboot.
The training I am doing is also not that "hard", you can see here the output of nvidia-smi:
Wed Nov 18 09:12:08 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 455.38 Driver Version: 455.38 CUDA Version: 11.1 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... Off | 00000000:01:00.0 Off | N/A |
| 18% 57C P2 78W / 280W | 9677MiB / 11176MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 505 G /usr/lib/Xorg 9MiB |
| 0 N/A N/A 532 G /usr/bin/gnome-shell 3MiB |
| 0 N/A N/A 1308 C python 9659MiB |
+-----------------------------------------------------------------------------+As you can see the memory is not even full, the GPU fan is at 18%, temperature is OK, power is low.
I really have no idea what is wrong, and any help would be very welcome.
Offline
I suspect a hardware issue
It could be
1. Temperature
2. Ram
3. PSU
and check in that order
you listed the temp of the gpu, but not the cpu? if you are sure nothing got too hot, check the ram with Memtest86 to be sure you dont have a defective stick
also this is very helpful
# previous boot, highest priority
journalctl -b -1 -p3 Last edited by ponyrider (2020-11-18 09:36:43)
Offline
I suspect a hardware issue
It could be
1. Temperature
2. Ram
3. PSUand check in that order
you listed the temp of the gpu, but not the cpu? if you are sure nothing got too hot, check the ram with Memtest86 to be sure you dont have a defective stick
also this is very helpful
# previous boot, highest priority journalctl -b -1 -p3
Trying your command, I only get some errors that happened at boot time, far away from when the computer rebooted:
Nov 17 09:22:43 auchentoshan kernel: x86/cpu: VMX (outside TXT) disabled by BIOS
Nov 17 09:22:44 auchentoshan kernel: bcma-pci-bridge 0000:05:00.0: bus0: No SPROM available
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:44 auchentoshan kernel: Bluetooth: hci0: BCM: firmware Patch file not found, tried:
Nov 17 09:22:44 auchentoshan kernel: Bluetooth: hci0: BCM: 'brcm/BCM20702A1-0b05-180a.hcd'
Nov 17 09:22:44 auchentoshan kernel: Bluetooth: hci0: BCM: 'brcm/BCM-0b05-180a.hcd'
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:44 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:44 auchentoshan kernel: bcma-pci-bridge 0000:05:00.0: bus0: HT force timeout
Nov 17 09:22:45 auchentoshan kernel: bcma-pci-bridge 0000:05:00.0: bus0: PLL enable timeout
Nov 17 09:22:45 auchentoshan kernel: b43-phy0 ERROR: FOUND UNSUPPORTED PHY (Analog 12, Type 11 (AC), Revision 1)
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: CPU SrcID #0, Ha #0, Channel #0 has DIMMs, but ECC is disabled
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Couldn't find mci handler
Nov 17 09:22:45 auchentoshan kernel: EDAC sbridge: Failed to register device with error -19.Offline
all that says it that you haven't installed whatever broadcom device you have, i dont think the EDAC messages are relevant. check your ram!
Offline
I don't understand why it says that the previous boot is "still running".
Do you have a prallel windows installation?
Offline
I don't understand why it says that the previous boot is "still running".
Do you have a prallel windows installation?
There is a W10 installation in another disk, which hasnt been used or booted in years. This is a computer that I only ssh in, no screen attached. grub boots straight to linux.
nevertheless, here is my fdisk:
Disk /dev/sda: 1.82 TiB, 2000398934016 bytes, 3907029168 sectors
Disk model: WDC WD20EARX-00P
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: gpt
Disk identifier: A27CEA4D-2950-45FB-9C13-EF648AFBC658
Device Start End Sectors Size Type
/dev/sda1 2048 923647 921600 450M Windows recovery environment
/dev/sda2 923648 1128447 204800 100M EFI System
/dev/sda3 1128448 1161215 32768 16M Microsoft reserved
/dev/sda4 1161216 1039831039 1038669824 495.3G Microsoft basic data
Disk /dev/sdb: 111.79 GiB, 120034123776 bytes, 234441648 sectors
Disk model: SanDisk SD8SBBU1
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 13338038-1275-414F-8A49-C9B8B1AB7721
Device Start End Sectors Size Type
/dev/sdb1 2048 115345407 115343360 55G Linux filesystem
/dev/sdb2 115345408 234441614 119096207 56.8G Linux filesystem
Disk /dev/sdc: 931.51 GiB, 1000204886016 bytes, 1953525168 sectors
Disk model: ST1000DM003-1ER1
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: gpt
Disk identifier: F7B1D413-343C-4F16-8BCA-647C515E4008
Device Start End Sectors Size Type
/dev/sdc1 2048 1953525134 1953523087 931.5G Linux filesystemOffline
all that says it that you haven't installed whatever broadcom device you have, i dont think the EDAC messages are relevant. check your ram!
Is there any way to know if the RAM is the one giving me problems?
The Deep Learning stuff I am doing doesnt use that much RAM, it mostly loads from disk. You can see from top:
top - 12:25:26 up 6:37, 3 users, load average: 4.70, 4.95, 4.96
Tasks: 253 total, 7 running, 246 sleeping, 0 stopped, 0 zombie
%Cpu(s): 39.9 us, 6.8 sy, 0.0 ni, 53.2 id, 0.0 wa, 0.1 hi, 0.0 si, 0.0 st
MiB Mem : 96488.8 total, 84517.3 free, 7680.5 used, 4291.0 buff/cache
MiB Swap: 0.0 total, 0.0 free, 0.0 used. 87678.0 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
38650 me 20 0 30.4g 4.0g 183416 R 85.4 4.3 0:02.57 python
38652 me 20 0 30.4g 4.0g 183444 R 85.4 4.3 0:02.57 python
38654 me 20 0 30.4g 4.0g 183428 R 85.4 4.3 0:02.57 python
38656 me 20 0 30.4g 4.0g 183444 R 85.4 4.3 0:02.57 python
38658 me 20 0 30.4g 4.0g 183436 R 85.4 4.3 0:02.57 python
38660 me 20 0 30.4g 4.0g 183440 R 85.4 4.3 0:02.57 python
7893 me 20 0 29.6g 5.0g 1.7g S 26.9 5.3 63:05.72 python (I have replaced my name with "me")
Offline
The disk doesn't matter - boot windows and ensure it's not hibernating (notably because of fast-boot, 3rd link in my signature)
Offline
The disk doesn't matter - boot windows and ensure it's not hibernating (notably because of fast-boot, 3rd link in my signature)
I see, but if windows hasnt been booted in 2-3 years, and until recently everything was going OK, I would assume windows is not the problem? Why would boot the system now after so long?
Offline
ACPI voodoo - there's no guarantee that this is the case but *iff* the BIOS believes there's a hibernated system, weird things can be expected.
You need to rule out every possible cause and de-hibernating windows is going to be quicker than eg. running memtest86 (a reasonable amount of cycles will be counted in days)
Offline
Pages: 1