On 10/12/2018 16.57, Per Jessen wrote:
Carlos E. R. wrote:
On 10/12/2018 15.04, Per Jessen wrote:
Carlos E. R. wrote:
On 10/12/2018 14.52, Per Jessen wrote:
Carlos E. R. wrote:
...
Notice I didn't say "give more resources", I suggested you reserve memory to make sure the GUI remains responsive.
Rather, it would be limiting some resource to FF, but which?
Memory. By reserving memory for the GUI, you are also limiting what other app can use.
I don't think that will help. Notice that the CLI was not responsive. Nothing was responsive. Giving more memory to the GUI (what process) will not help with the CLI.
After you start the GUI, all subsequent processes are in the same cgroup. That way you will have e.g. konsole remaining responsive. I think we ought to have this config by default, at least on machines with 4Gb or more.
No, CLI over ssh is not on that group.
Correct, it isn't GUI but presumably you wouldn't need the remote access when the GUI remains responsive.
I prefer to don't change anything yet till we really know what is happening, and for that I need to know what command to issue on the CLI that gives the same information as the sysrq-m that Andrei Borzenkov suggested. I'm running the laptop with the same workload that crashed it, with 4 ssh sessions opened from my desktop computers, both left permanently running in the hope to catch it. So far, no luck - but I don't know that command, so if it happens I can do nothing.
top - 03:40:14 up 4 days, 1:38, 5 users, load average: 0.08, 0.11, 0.14 Tasks: 287 total, 1 running, 286 sleeping, 0 stopped, 0 zombie %Cpu(s): 4.6 us, 1.9 sy, 0.0 ni, 93.2 id, 0.1 wa, 0.0 hi, 0.1 si, 0.0 st KiB Mem : 3934240 total, 834848 free, 2485356 used, 614036 buff/cache KiB Swap: 6289412 total, 4982200 free, 1307212 used. 950024 avail Mem
-- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)