Page 1 of 3 123 LastLast
Results 1 to 10 of 25

Thread: Laptop cpu suddenly went to 99%, now can boot but not login.

  1. #1
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Laptop cpu suddenly went to 99%, now can boot but not login.

    Hi. This charming new problem occurred contemporaneously whilst i was working on an unrelated thread several hours ago today https://forums.opensuse.org/showthre...76#post2845376 , from which i now quote my Postscript:
    ...in Lappy's YaST FW settings, Lappy suddenly became severely laggy to all inputs, & cpu went to 99%. Whilst i could not get to see what was eating the cpu, i know that it was not the btrfs maintenance [balancing] cron job, as that runs on Tuesday nights [& it's my Sunday afternoon now]. Eventually with no improvement i had to Ctrl-Alt-Backspace to kill the session, but that did not help as once logged in again it was still overloaded. I tried to reboot but it was ignored. I REISUB'd, & thereafter discovered the extent of the apparent disaster. I can boot, "apparently" unlock my encrypted /home at the screen for that as usual, but then cannot actually login to my Plasma or IceWM desktop anymore, as me or my other user. I tried multiple Snapper Rollbacks, but all gave no change to the preceding symptoms. I rebooted from Live media then ran
    Code:
    sudo btrfs check --repair /dev/sda3
    as previously worked for me in Tower [https://forums.opensuse.org/showthre...1#post2840121], but all the preceding symptoms [boot, "decrypt & mount", login just bounces back to login screen] continued.
    Given that notwithstanding the "apparent" success of unlocking my Lappy's crypto_LUKS /home [in that no error msg appears, & all seems to proceed as normal from there to the subsequent desktop login screen], the observation that every login attempt invisibly fails [ie, no error msg] & instead the login screen simply keeps reappearing, rather than proceeding to my desktop, makes me wonder if somehow whatever suddenly drove the cpu to 99% & initiated this mess might have somehow damaged my encryption system, ie, maybe it's not really decrypting & mounting any longer...?

    Does anyone pls have any suggestions, or need i prepare myself for a full reinstallation?

  2. #2
    Join Date
    Jun 2008
    Location
    Groningen, Netherlands
    Posts
    16,725
    Blog Entries
    13

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    At the login screen, hit Ctrl-Alt-F2, and first try to login with username and password. If that doesn't work try to login as root, find out whether /home is mounted. It seems that the culprit is outside of the system's snapshot. Any backups of /home ?
    ° Appreciate my reply? Click the star and let me kow why.

    ° Perfection is not gonna happen. No way.

    http://en.opensuse.org/User:Knurpht
    http://nl.opensuse.org/Gebruiker:Knurpht

  3. #3
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Thanks. Yes i can, kinda sorta, login at tty2, but... unpleasant message there:
    You have mail.
    -- gooeygirl: /home/gooeygirl: change directory failed: No such file or directory
    Logging in with home = "/".
    Am i correct to interpret this as implying that the system cannot find my /home, presumably because the decryption & mounting failed?

    Still in tty2 i logged in as root, via "su -". Transcribing here [am writing this in Tower, whilst looking at Lappy next to me], i changed to / directory, then ran "df -h". Neither /dev/sda4 [encrypted Swap] nor /dev/sda5 [encrypted /home] is listed [ie, i presume = not mounted]. To restate that maybe more accurately, when i ran "df -h" just now in my Tower, one of its entries is:
    Code:
    /dev/mapper/cr_ata-Samsung_SSD_850_EVO_250GB_S21MNSAG105383J-part3  160G   37G  124G  23% /home
    
    . There is NO equivalent of that appearing in the output in Lappy's tty anymore, & of course there should be.

    Prior to finding your reply, i did a quick little experiment on the unhappy Lappy. I booted it again then at the screen where i enter my password to supposedly decrypt & mount my /home, i deliberately entered a wrong string, to see how it would react. It immediately recognised that my password was wrong & prompted me to try again. This time i deliberately entered another wrong one, & it again saw my error & re-prompted me. This time i did enter my correct password, after which it proceeded on as normal to the standard TW login screen [where sadly i now always get stuck in groundhog day]. To me this experiment implies that TW is still aware of my correct encryption password... but for some reason is not correctly using it.

    In a few minutes i am going to try to take a picture of Lappy's tty2 screen showing the "blkid" & "df -h" outputs. If the picture is able to adequately show the text legibly i'll then attempt to upload it here in a new post.

  4. #4
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    blkid picture, Lappy, tty2:


  5. #5
    Join Date
    Aug 2010
    Location
    Chicago suburbs
    Posts
    9,164
    Blog Entries
    3

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Quote Originally Posted by GooeyGirl View Post
    Am i correct to interpret this as implying that the system cannot find my /home, presumably because the decryption & mounting failed?
    Yes, probably something like that.

    However, if it cannot find "/home", I would have expected it to boot into emergency mode. But if "/home" is okay, but you have an encrypted personal home directory, then it would probably do what you saw if something went wrong.

    To restate that maybe more accurately, when i ran "df -h" just now in my Tower, one of its entries is:
    Code:
    /dev/mapper/cr_ata-Samsung_SSD_850_EVO_250GB_S21MNSAG105383J-part3  160G   37G  124G  23% /home
    
    . There is NO equivalent of that appearing in the output in Lappy's tty anymore, & of course there should be.
    Okay. But that's for an encrypted home partition. For your laptop, I'm confused as to whether you have an encryped home partition, or an encrypted home directory.
    opensuse Leap 42.3; KDE Plasma 5;
    opensuse tumbleweed; KDE Plasma 5 (test system);

  6. #6
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    df -h picture:



    Home & Swap are not mounted; bwaaaah.

    Btw, i forgot to answer your other question -- yes i do have backups, so if i have to go nuclear & do a full reinstallation, whilst being a major PITA it's not actually catastrophic. What i am more worried about though is WHY this might have occurred, & if it has any implications for a pending similar event on my Tower [= my primary pc]. Furthermore, until/unless the root cause is known, it now makes me wonder anew whether TW is right for me. That it happened at all is worrisome, & the discovery that even Snapper Rollbacks could not help, now potentially nullifies what up til now has been my unwavering support of BtrFS & Snapper with TW. However, without a proven cause, all that is just futile speculation.

  7. #7
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Quote Originally Posted by nrickert View Post
    Okay. But that's for an encrypted home partition. For your laptop, I'm confused as to whether you have an encryped home partition, or an encrypted home directory.
    Thank you, & i apologise. Despite my attempts over the months to sort this out in my head, & despite Henk's previous valiant attempt to help me understand, i seem too dimwitted to get my terminology correct.

    Here are 2 pics i took during my original installation of TW on Lappy earlier this year; do they fully & unambiguously answer your question?




  8. #8
    Join Date
    Aug 2010
    Location
    Chicago suburbs
    Posts
    9,164
    Blog Entries
    3

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Okay, so that's an encrypted partition for "/home" (and another for swap).

    One possibility: something has changed in the partition naming. I suggest you check "/etc/crypttab", and see whether the entries still identify disk partitions. The second column is probably of the form "/dev/disk/by-id/some-funky-string".

    Do an "ls -l" on whatever that shows. See if there is a device there (or a symbolic link to a partition).
    opensuse Leap 42.3; KDE Plasma 5;
    opensuse tumbleweed; KDE Plasma 5 (test system);

  9. #9
    Join Date
    Jun 2017
    Location
    Australia
    Posts
    582

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Thanks. Yes "encrypted partition" was what i wanted to say, but i hesitated in case i got it wrong & then caused mass confusion.

    Back in Lappy's tty2, as root, i accessed "/etc/crypttab" & viewed its contents with nano. Fortunately in my Tower [my primary pc] i keep records of both Tower's & Lappy's customised [by me] system files, just in case... Here's Lappy's one, as customised by me back in July soon after installation:
    Code:
     #cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part4 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part4 none       none
    cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part4 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part4 none       luks,allow-discards
    #cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 none       none
    cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 none       luks,allow-discards
    I carefully, visually, compared both active strings as above, to what nano is displaying now --> still identical.

    You said;
    Do an "ls -l" on whatever that shows
    Sorry but i don't understand. How/where do i do that? Surely before i could do that, the device must be mounted [or decrypted, or "available"; sorry, but i still cannot master the essential concept here], but as my second pic [df -h], above, showed already, there is NO...
    Code:
    /dev/mapper/cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5
    , hence it is not mounted / available [?]

    For comparison, fyi here is my Tower's corresponding string from its "/etc/crypttab" for its encrypted partition on which /home is mounted:
    Code:
    cr_ata-Samsung_SSD_850_EVO_250GB_S21MNSAG105383J-part3 /dev/disk/by-id/ata-Samsung_SSD_850_EVO_250GB_S21MNSAG105383J-part3 none       luks,allow-discards
    ...& here is that same device correctly listed in Tower's df -h:
    Code:
    /dev/mapper/cr_ata-Samsung_SSD_850_EVO_250GB_S21MNSAG105383J-part3  160G   37G  124G  23% /home
    As that earlier df -h picture showed for Lappy, its necessary
    Code:
    /dev/mapper/cr_ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5 /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part5
    ...is simply... missing, somehow. But it's not like Lappy's encrypted sda4 [=swap] & sda5 [=/home] partitions have "vaporised" & "ceased to be", coz my first picture above [blkid] still lists them both.

    I'm very confused about what happened, out of the blue, yesterday... & what if anything i can do now.

  10. #10
    Join Date
    Aug 2010
    Location
    Chicago suburbs
    Posts
    9,164
    Blog Entries
    3

    Default Re: Laptop cpu suddenly went to 99%, now can boot but not login.

    Quote Originally Posted by GooeyGirl View Post
    Sorry but i don't understand. How/where do i do that?
    After you have booted, and logged in at the console (as root or as yourself), do:
    Code:
    ls -l /dev/disk/by-id/ata-SAMSUNG_SSD_PM810_2.5__256GB_S0N4NEAZB01960-part4
    It doesn't matter if the disk is mounted and decrypted. This just checks whether the kernel is still recognizing that device-id.

    Maybe, also do:
    Code:
    ls -l /dev/disk/by-id  |  grep sda4
    I'm guessing that something has changed in how the device-id is calculated.
    opensuse Leap 42.3; KDE Plasma 5;
    opensuse tumbleweed; KDE Plasma 5 (test system);

Page 1 of 3 123 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •