20.04.2 is where I am having problems.
5.4.0-65-generic and 5.4.0-42-generic are the kernels I have, am guessing that the latest one was part of the upgrade to 20.04.2.
Discover won’t run in either of those, but it works both of the respective recovery modes.
The version of Discovery is 5.18.5
Is this what you meant?
Thanks
Yep, and that’s what I have. I’m thinking hardware more and more. You should try the other tests I suggested (Kubuntu, 20.10).
Have just tried Kubuntu , never used it before, but its very swish. 20.10 Discover 5.19.5 opens and then offers up 207 updates. I’m currently downloading Lubuntu 20.10, but will install Kubuntu on a partition in the meantime and then update it to see what happens.
Okay after allowing Kubuntu to install updates I got the same problem I think. Segmentation Error, although the installer also crashed at something to do with grub and so did the updater. It then failed to restart/logout the friendly menu or something crashed. Then after reboot through hard reset, it didn’t quite load up KDE (is that the desktop?) got a black screen and a mouse cursor.
Going to try and run the install again but without updates or 3rd party plugins. Although I am thinking I have problems with either the RAM, the CPU or both. I see errors on boot to do with hardware error - cpu 1, processor 2 and something to do with timer and aio.
Kubuntu is pretty fantastic. Given it’s driven by Qt, we’re close relatives in the Ubuntu family.
Unfortunately, your problem reminds me of a favorite motto: Linux can do a lot of amazing things, but it can’t fix broken. I admit that the issue might not be that something is damaged… but it might; I’ve seen it many times before. The only way to test this is usually costly: swapping out components.
At the least, I can only imagine one explanation for the difference between the results you’re having and what others are. It has something to do with your particular hardware, which is the one remaining variable we can’t easily test for (even were we to have the exact same model).
I can’t say why it seems that it’s Discover in particular that’s triggering the problem. I don’t think it does anything particularly remarkable as far as system calls (or anything else) are concerned. It makes me wonder what other software might produce the same symptoms. What other software stores? What about graphical package managers like the aforementioned Muon? Or anything else? What software has worked?
Regardless, the hardware errors are pretty telling, as they indicate cases where the hardware is not operating as it should. In fact, with specific detail we could probably figure out the exact problem. Look here and you’ll be able to see a case where it was obviously the RAM while here is another where it was determined to be the CPU and backed by a similar tale.
To explore the idea of the hardware issues, I’d probably dig through system logs and see if anything in particular jumps out. You could even do journalctl -k | pastebinit (which limits things specifically to kernel logs) and share the resulting URL with us and we can have a look, too. Sometimes a very careful read can lead one to clues as to where the problem might lie.
Running a memtest is cheap, so I’d probably do that as an initial diagnostic.
I’ll set a memtest running in a bit, but the output from journal was:
https://paste.ubuntu.com/p/JqVPJBHhcm/
I noticed the following highlights (you might notice more than I):
8254 timer not connected to io-apic
EISA cannot allocate resource…
Unstable clock
AMD64 ECC Disabled
ata3 soft reset failed
I did a bit of reading and some of these point towards setting acpi=off / radeon.modeset = 0 noapic - so thinking they are related to the graphics.
I see the same error as Discover for Firefox, except it still runs. I tested Muon and that worked fine, for checking for updates and downloading / install some packages. Currently as I type this I have Minetest running between 20 - 30 fps, Firefox open with 7 tabs, Quassel IRC, Muon Package Manager, Terminal and a QT File Manager Window open at the same time. None of these seem to have any issues, I also opened up Libre Office quickly and that seemed to start fine. I am in two minds at the moment about whether to revert to 20.04.1 or just stick with 20.04.2 and use Muon; I do have a work around to use Discover in recovery mode and its only use is really to be able to read more about the packages and some comments by users I suppose.d
I have also been getting a hardware warning / machine check error at boot too, this is probably the most concerning, do these get logged somewhere too? Or are they in journal somewhere? I have looked at my CPU and I have some RAM stashed somewhere, but upgrading / replacing those components wouldn’t be too expensive beyond the time taken to do it.
This is the reading I have been doing around the warnings / errors in journal:
https://ubuntuforums.org/showthread.php?t=2283568
https://www.linuxquestions.org/questions/linux-kernel-70/kernel-fails-to-assign-memory-to-pcie-device-4175487043/
https://ubuntuforums.org/showthread.php?t=949593
Feb 21 09:07:17 alfie-lubuntu kernel: firefox[1105]: segfault at 0 ip 0000000000000000 sp 00007ffd572bdf18 error 14 in firefox[55b7eb22c000+b000]
Feb 21 09:07:17 alfie-lubuntu kernel: Code: Bad RIP value.
Feb 21 09:12:25 alfie-lubuntu kernel: plasma-discover[1625]: segfault at 0 ip 0000000000000000 sp 00007ffe114de2d8 error 14 in plasma-discover[55beea80c000+14000]
Feb 21 09:12:25 alfie-lubuntu kernel: Code: Bad RIP value.
Feb 21 09:45:27 alfie-lubuntu kernel: plasma-discover[2131]: segfault at 0 ip 0000000000000000 sp 00007ffec6088238 error 14 in plasma-discover[5611cf5db000+14000]
Feb 21 09:45:27 alfie-lubuntu kernel: Code: Bad RIP value.
Feb 21 09:46:07 alfie-lubuntu kernel: soffice.bin[2166]: segfault at 0 ip 0000000000000000 sp 00007ffdc02a4698 error 14 in soffice.bin[5652c38c4000+1000]
Feb 21 09:46:07 alfie-lubuntu kernel: Code: Bad RIP value.
Feb 21 09:46:14 alfie-lubuntu kernel: soffice.bin[2195]: segfault at 0 ip 0000000000000000 sp 00007ffdfb967b68 error 14 in soffice.bin[55b39a11c000+1000]
Feb 21 09:46:14 alfie-lubuntu kernel: Code: Bad RIP value.
I took a snip from the paste you provided. It actually looks like you had 3 different programs, Firefox, Discover, and LibreOffice, that are having a segfault issue. You’ll notice that there is a “Bad RIP value” right after the segfault. A “RIP” is a memory Register Instruction Pointer. When you have a Bad RIP value that means the memory at the pointer is not executable for some reason. It might mean that the memory is having an issue or the contents pushed to the memory from the file system is bad. So my 2 suggestions are to look at the memory test and to take a look at the drive to make sure there is no drive corruption or SMART errors.
Thanks I didn’t notice the Libre Office error. The memory test has been running for 6 hrs 30 minutes and says Pass 3 Errors 0. Does this mean it has completed three full tests? Its at Test #8 but I feel sure when I checked earlier it was on Test #10. Does it just keep going forever?
Update.
I let memtest86+ run through a fourth time and there were no errors reported.
Then I ran a short smartctl test and it suggests that the hdd is old, very old and in some places pre-fail… or at least that’s what I could make of it. Said it has been run for 52 days and 4 hours in total. Is that a lot?
I do have a 500gb drive somewhere I could test and see if is any better. Tbh most of my drives are probably quite old.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.