Abrowser very frequent crashes
Abrowser version 20.0 in Trisquel 6.0 - 64bit version
I have newly installed this system today, and I love it.
But I have this issue with abrowser, which crashes constantly,
I can hardly use it few minutes at the time.
Also some gui freezings happen. I have reinstalled the browser and the problem persists. I wonder if anybody has similar problems, and possible fix.
Try running the browser from a terminal to see if you get an error message.
When you say GUI freeze, do you mean xserver freezes? How do you unfreeze it? Does sysrq work if nothing else seems to? https://en.wikipedia.org/wiki/Magic_SysRq_key
Yes, this is really mystical. I have run abrowser from terminal, and now there are no error messages nor crashes anymore. I will post later if this problem reappears. Maybe running abrowser first time from the panel creates some problems. Now I ran it also from panel and no problem.
regarding GUI:
all windows freezed couple of times (but not anymore), usually when installing new packeges from synaptic. I could usually move mouse, and it was unclear if processes were finished or not. I resolved problem by brute force restart.
I will report more, if something unexpected happens. At the moment I am so overwhelmed with Trisquel.
It happened again while I ran it through the terminal,
and I got an error saying:
"Segmentation fault"
When I reattempted to run abrowser I was asked if I would like to reset to factory version or safe mode. I have reset it, and so it is now until the next crash, although it is less often.
Maybe it is just the small instability of the system. Also I was thrown out of my user session and I had to reload my indicator applet.
So far I got this error:
Gtk-Message: Failed to load module "canberra-gtk-module"
Segmentation fault
This module was not installed by default. I have installed it now, maybe this was the problem.
I am a member!
"Segmentation fault" means memory management problems. I would install "memtest86+" reboot the system and choose MemTest in GRUB menu; if it accuses no errors after completing a test cicle, I would bet there is some bug in your software -- have you checked the integrity of the installation media (if any) before installing the system?
P.S.: if it shows errors, anotate the address ranges (something like "0x01234567,0xfefefefe"), write them into the "GRUB_BADRAM=" field in /etc/default/grub and run "$ sudo update-grub": your kernel will not use the bad addresses again -- of course, others may appear.
Thank you! I think it is almost with certainty that there is some problem with my memory, although it is not so old (2 years, brand Patriot).
First I could not boot into the memtest86+ from GRUB (neither a console version), and neither after I switched into AHCI mode for SATA.
I have two memory chips, and so I unplugged one, and tried again, I could not boot neither. Then I unplugged both of them and found lot's of dust behind one of them, which was thrown from the CPU ventilator. I have then cleaned with vacuum cleaner, left memory unplugged few minutes and put it back again.
Now system seems more stable, nothing crashed so far, but I am still observing.
I wonder if I should buy a new memory soon, and why it gets destroyed so easily.
The problems persist. Now the whole system collapses. There are some error messages like: missing free microcode, running script... etc. when booting.
I would like to run memtest86+ but I can not boot it. I do not know where to look for a solution for that. I changed SATA mode from IDE to AHCI back and forth, I have tried with one memory chip at the time and nothing so far.
I am prepared to buy a new memory if necessary, but I do not know is it really that which will solve the problem. Or is it sufficient if I write those addresses to "GRUB_BADRAM="? Will that make my system inefficient? Please help!
The dust could be a problem. Missing microcode probably not so much.
>I would like to run memtest86+ but I can not boot it.
How come?
That badram parameter will tell your computer to not use the damaged regions of memory but first you'll have to locate them, if any. http://www.gnu.org/software/grub/manual/grub.html#badram
I am a member!
Indeed, how come? What happens?
Many GNU/Linux distros embed MemTest86+ within their LiveCD images; I believe Trisquel does that too; if you can boot from a LiveCD (or USB things) you could run MemTest86+ from there.
Yes, memtest86+ was from trisquels repository, and I tried without success many other live distributions. Until I found Fedora 16 and memtest86+ v.4.20, it booted. Then I run 3 passes of test, without ERRORS. I do not know what to do now. Even other programs, like Synaptic sometimes crash. Specialy after using youtube and gnash. Strange!
I am a member!
The version of memtest86+ in Trisquel repos is 4.20 also.
I would try reinstalling the core components (I believe everything under trisquel-minimal, including its dependencies' dependencies and so on); if that doesn't work, I would try reinstalling the system again.
But are you sure that the media you used to install Trisquel, if any, was 100% OK? You can test it from its boot menu.
I can't help with investigating your machine, since I know almost nothing about the hardware market.
SOLVED !!
OK. So, in my concrete case it was overheating of CPU, i.e. hardware problem. I used supplied CPU-fan, and got often complete system crash with "tainted kernel" message, and peculiarly those abrowser segmentation errors. This I have resolved now with water cooling of CPU.
Trisquel works like a charm, and YES, of course I check the media before I install.