You are not logged in.
Pages: 1
my computer randomly stops responding
like if im working with my browser suddenly pages don't load
and if i open a terminal the terminal opens fine but the shell doesn't start
my setup
cpu: Intel Core i3-4130
GPU: GeForce GTX 660
ram: 16 GB
1 120GB SSD for /
3 10TB HDD in raid5 for /homei think i tracked it down to mdadm.
but i don't know where to look from here.
if i change to a different tty and login as root i can do normal cli things
but as soon as i try to read/write a file in /home/vista that tty also stops responding
note* other folders in /home but not /home/vista work fine
if i look at top i can see that the program that stopped responding had a status D
i believe that's for using Disc
journalctl talks about how the raid timed-out
but i don't know if there is more information in there
the whole journal
i tried to check the discs with smartctl
but that gives a lot i dont know how to interpertate
the best i can get out of that is
SMART overall-health self-assessment test result: PASSEDsmartctl output
sda raid5 hdd
sdb raid5 hdd
sdc raid5 hdd
sdd root sdd
/etc/mdadm.conf is mostly commented out
there is only 2 lines with "DEVICE partitions" and "MAILADDR [my-mail]"
so i think there can be 2 things that don't work
mdadm or something involved with that
a HDD's is failing
isn't mdadm made keep working in case something is wrong with a drive ?
and how can i find out witch drive is causing the trouble ?
please keep in mind its very random when it decides to stops responding it can be 2 hours or 2 days.
so things to try or test while its stuck can take a while
and when changing settings i don't if it worked or not (unless we find a way to force the problem)
im not very experienced with this
im trying to provide as much info as possible
if there is something missing please ask (and maybe tell me how to get it)
Last edited by vista_narvas (2018-12-29 16:35:28)
Offline
Hello,
First and foremost, D stands for interruptible sleep, see https://idea.popcount.org/2012-12-11-li … ss-states/. This means the process is hung in a system call, which could be IO syscall. Nevertheless, look at dmesg and run smartctl -a /dev/sdXx smartctl -a /dev/sdX on each of your disks to determine their health.
EDIT: In fact, run smartctl -t short /dev/sdaX, or alternatively smartctl -t long, on each disks.
Regards
Last edited by bugsmanagement (2018-12-28 23:34:41)
Offline
Hello,
First and foremost, D stands for interruptible sleep, see https://idea.popcount.org/2012-12-11-li … ss-states/. This means the process is hung in a system call, which could be IO syscall. Nevertheless, look at dmesg and run smartctl -a /dev/sdXx smartctl -a /dev/sdX on each of your disks to determine their health.
EDIT: In fact, run smartctl -t short /dev/sdaX, or alternatively smartctl -t long, on each disks.
Regards
i totally forgot about dmesg
this is the output after boot
i did run smartctl -t short on all discs but long is going to take 800 minutes
and this is the output of smartctl
sda raid5 hdd
sdb raid5 hdd
sdc raid5 hdd
sdd root sdd
Offline
You should capture dmesg when the problems occur, as the supplied dmesg log doesn't contain anything glaring, neither does the smart logs. It could be intricacies with raid that I don't know much about.
Offline
ok i will make a copy of dmesg the next time
but that could take a day or 2
i will post it here as soon as that happens
Offline
The journal contains kernel messages. Systemd#Journal Could you locate and post all the kernel messages from a boot with the issue?
Offline
The journal contains kernel messages. Systemd#Journal Could you locate and post all the kernel messages from a boot with the issue?
i dont understanding the question
there are no problems during boot
only after some random amount of time
im at work right now
when im home ill see what i can find
Offline
I asking you to search through the journal entries for previous boots. Find the one containing
Dec 28 16:33:07 Narvas kernel: INFO: task md254_raid5:371 blocked for more than 120 seconds.Then post all the kernel messages from that boot. The following should provide the negative index of the boot covers that time period.
journalctl --list-bootsThen to get kernel messages from that boot substituting the negative index for #
journalctl -b # -kOffline
Try setting traditional (non-ssd) HDDs to BFQ scheduler (instead of mq-deadline).
https://wiki.archlinux.org/index.php/im … _scheduler
sys2064
Offline
I asking you to search through the journal entries for previous boots. Find the one containing
Dec 28 16:33:07 Narvas kernel: INFO: task md254_raid5:371 blocked for more than 120 seconds.Then post all the kernel messages from that boot. The following should provide the negative index of the boot covers that time period.
journalctl --list-bootsThen to get kernel messages from that boot substituting the negative index for #
journalctl -b # -k
the result of journalctl -k
Last edited by vista_narvas (2018-12-29 17:37:04)
Offline
ok so my pc stopped responding again
this is re result of dmesg after it stopped working and this is from after the reboot
and the journal -k of that boot session
the dmesg more or less contains the same info as the journal
and i dont see anything special before it happens
Offline
Pages: 1