Ryzen build locks up or freezes.
I've been running Fedora 28 with a Ryzen 3 1300x and it often freezes and is unresponsive within 1 week or sooner. A reset is required. I've experimented with Manjaro and haven't had the same problem occur. However, a build of Majaro that I ran in April for more than one month may have frozen too(vague memory). Point, Fedora builds lockup reliably and quite quickly. Anyone have clues here?
update: My Manjaro build SSD froze after nearly two weeks. Restarted the Fedora 28 SSD and it is freezing nearly daily. Whoa.
Have you checked for recent firmware updates to the Motherboard?
You may also want to compare kernel versions of Manjaro and Fedora.
After a freeze, are you able to get kernel logs from the previous boot using
journalctl -k -b -1?Yes, motherboard firmware is latest, the Fedora kernel is newer than Manjaro but not by much. And no, I can't access the logs during a freeze, even if I knew what to look for( I'm not that sophisticated) Will investigate. See the update on post. edit: I ran your command, and am very much in the dark. RH
Do you get any output from the command? If not, what is the output of
journalctl -p 3 -r -n 15 -b -1after you reset your computer? This command looks at the last 15 lines of error (or higher) log entries from the last boot. (you can see other options for journalctl from https://www.commandlinux.com/man-page/man1/journalctl.1.html )Here is the output from your latest command:
Tue 2018-06-05 10:33:57 CDT, end at Mon 2018-08-06 16:04:24 CDT. -- Aug 05 12:02:11 localhost.localdomain kernel: sp5100-tco sp5100-tco: Watchdog hardware is disabled
Mean anything to you?
RH
I ran across a Arch wiki post dealing with the sp5100-tco ---suggested creating a /etc/modprobe.d/sp5100-tco.conf and blacklist.. ...will monitor.
The watchdog hardware (if you're not familiar) will reboot the computer if it locks up after a given time. The sp5100-tco has been often mis-identified and the module loaded as an intel i2c chip. Does
lsmod|grep i2c-piix4give any output? If so, look to see which driver the watchdog has loaded (lspci -kshould do the trick, though you may have to sort through the output of that.) I've run into this same issue with an earlier AMD chip and an Asus Motherboard. Blacklisting the i2c-piix4 module solved my problem.No, the grep gave no output. However the lspci -k revealed this: Kernel driver in use: piix4smbus Kernel modules: i2cpiix4, sp5100_tco
should I add the i2c...to the blacklist.
Over night the machine clock stopped at 6:30am assuming the freeze was then, and a new journalctl command revealed:
Logs begin at Tue 2018-06-05 10:33:57 CDT, end at Wed 2018-08-08 08:34:28 CDT. -- Aug 07 16:33:33 localhost.localdomain spice-vdagent[1575]: Cannot access vdagent virtio channel /dev/virtio-ports/com.redhat.spice.0
Thanks for your ongoing help. RH
Update: With the i2cpiix4, sp5100_tco blacklist in effect my computer machine has been running for 2 days with no freeze...hoping that has done it.
update; I'm about to proclaim this a fix for my lockup/freezing problem. I've been booted up for 8 days on my computer machine and no freeze. In fact I'm sensing, maybe imagining, that the mouse no longer shows a slight hesitancy that was there (maybe) before.... Just a subtle thing. Fingers crossed. Will leave running for a couple of days more. Before the blacklisting, 10 days was about as long as I could go without a lockup.
Good! Glad blacklisting the watchdog seems to have fixed the issue!
Thank you very much Mr K. Should I leave this posting for others our delete?
I'll add it as an answer, if you could accept it , that would be great!
We've been a little premature. Froze last night, discovered this morning. Ouch. Trying this: In a Manjaro forum post a person inserted 'modprobe.blacklist=i2cpiix4' in the /etc/default/grub file as such: GRUBCMDLINELINUXDEFAULT="sysrqalwaysenabled=1 modprobe.blacklist=i2c_piix4 radeon...etc.
giving that a try. RH