Should I post a freeze I can pin down?

  • This topic has 7 replies, 4 voices, and was last updated Jan 30-10:46 pm by rayluo.
Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #50999
    Member
    roland

    Dear Friends,

    Using 64-bit 19.3 and kernel 5.8.16 I can quite simply create and recreate a freezeup or workspace lockup whenever I want, and can do so for further investigation, if thought to be of sufficient importance by the moderators. I can even show there is a simple way to get past it despite the workaround being somewhat inconvenient. I thought it may give insights into patch(es) that might be appropriate to design and apply to the kernel. Or maybe is isn’t a kernel problem at all?

    Please advise, and if thought worthwhile then please suggest what diagnostics would be appropriate to take before the freezeup occurs. Naturally after the freezup no further diagnostics are possible. Please be aware that I do not know how to take a memory dump on a PC.

    Initially I had decided to wait for a later version of the 5.8 kernel, because for hardware reasons I cannot regress to any lesser kernel version such as 4.9.xxx.

    #51021
    Member
    skidoo
    Helpful
    Up
    0
    :D

    worthwhile? Sure. We all learn from collaboratively solving puzzling problems.
    If you describe the repeatable steps leading to the problem, might find out that someone else has noticed the same (someone who perhaps only encountered the problem on one occasion, or didn’t connect the dots and recognize the repeatability…)

    > I do not know how to take a memory dump on a PC.

    S’okay.
    How many of us here would even “know what to do with” a submitted coredump datafile ?

    #51033
    Forum Admin
    rokytnji
    Helpful
    Up
    0
    :D

    dmesg has always been my friend in these type of situations
    also sensors-detect .syslogd even.
    Good luck

    Sometimes I drive a crooked road to get my mind straight.
    Not all who Wander are Lost.
    I'm not outa place. I'm from outer space.

    Linux Registered User # 475019
    How to Search for AntiX solutions to your problems

    #51056
    Member
    roland
    Helpful
    Up
    0
    :D

    I’ m not now inclined to post this in detail until I have thought about it more, as it is definitely not a kernel problem, much more likely an application problem, Chromium to be exact or maybe Facebook. But here is the outline of what is happening.

    It occurs in Facebook when a jpg image is enlarged by selecting the image with the pointer, then left click which makes it current, so it can be saved in a folder with right click and select required folder, then save. The jpg is saved correctly but no further activity is possible, the application becomes locked on the current workspace.

    However if I start another instance of Chromium on another workspace, and navigate exactly as before to the same jpg, I can continue as if there has been no freezup. If I then switch back to the original workspace and delete the entire instance of Chromium, I have worked round the freezup, and the new instance of Chromium continues to function ok.

    I have tried this on 2 PCs running markedly different kernels (4.9.xxx and 5.8.xx) but both 64bit cpus, and exactly similar behaviour takes place therefore I withdraw any suggestion of kernel involvement and accuse the application.

    I have been using Chromium for several years now and have not observed this behaviour until recently, from say 2 or 3 weeks since.

    Any observations gratefully received.

    #51079
    Forum Admin
    rokytnji
    Helpful
    Up
    0
    :D

    Hmmm. Seems the image you downloaded is screwing with settings in home folder for Chrome browser.
    Coming from facebook.
    Not a big surprise though.

    That is my guess . Was also wondering. You know you can launch browser from terminal. Open a Htop session in another terminal.
    All before the freeze. Fill up the screen 1/2 and 1/2 and hope when it freezes.
    You get some pertinent info.

    Sometimes I drive a crooked road to get my mind straight.
    Not all who Wander are Lost.
    I'm not outa place. I'm from outer space.

    Linux Registered User # 475019
    How to Search for AntiX solutions to your problems

    #51102
    Member
    skidoo
    Helpful
    Up
    0
    :D

    The jpg is saved correctly but no further activity is possible, the application becomes locked on the current workspace.

    visit the Chromium “preferences” interface and look for a setting mentioning
    “File Open/Save”
    The newer versions of both firefox-based and chromium-based browsers
    are eager to have you use (eager, as evidenced by the shipped default settings)
    their own (powered by browser component) file open/save dialog.
    The setting will be a toggle, labeled someting like “use the native…(or not)”.
    Try changing (toggling) that setting, and retest.

    A jam-up like you’ve described may be specific to (y)our exact installed version of some gtk library, or other component ~~ so might be difficult for upstream developers to ever notice (which explains how it managed to slip through as a “bug”) and might disappear, resolve itself, when a system update installs a newer version of the affected component.

    #51139
    Member
    roland
    Helpful
    Up
    0
    :D

    Thanks for all responses.

    The reason I was cautions about posting or not was I am ever aware of my own shortcomings, and this may be a case when I should have thought it through better.

    I feel it is a case where the destination folder is located on a volume not mounted. Some applications do not attempt to mount the volume in such a case, spaceFM I think is an example. Others will perform the mount if they can, Rox-filer for example. But in nether case do they hang if the save cannot complete. But Chromium simply hangs, perhaps endlessly leaving the workspace apparently frozen. I was clearly mistaken when I said the save had completed.

    I intend to check this in detail step by step including looking at the automount setting for volumes which at present is not ticked. I am using disk-manager to manually handle mounting tasks, and had overlooked having distributed various data folders across 8 volumes which are unmounted until I specifically mount them.

    Once again, I think this is not a problem at all simply lack of experience! Sorry for the waste of everybody’s time!

    #51211
    Member
    rayluo
    Helpful
    Up
    0
    :D

    The reason I was cautions about posting or not was I am ever aware of my own shortcomings, and this may be a case when I should have thought it through better.

    I feel it is a case where …

    I intend to check this in detail step by step …

    Once again, I think this is not a problem at all simply lack of experience! Sorry for the waste of everybody’s time!

    Don’t feel bad, Roland. We were all inexperienced at the beginning. 🙂

    With regard to the more general topic of how/when to post a potential issue, the software industry tends to favor step-by-step REPRODUCIBLE instructions. I’ve been at both sides of issue reports. The beautify of such a method is, it may seem cause more time on the reporter at first, but that could sometimes lead to a successful self-troubleshooting; for the rest of the time, the developer can follow up more efficiently. Some developer would even go ahead and put such guidance into their project repo website’s issue template.

    What you’ve been doing, is exactly the process of the self-troubleshooting. And you posted some messages during the process. I think that’s fine, considering this is a general purpose forum, not strictly a repo’s issue reporting template. 🙂 Don’t feel bad, my friend.

Viewing 8 posts - 1 through 8 (of 8 total)
  • You must be logged in to reply to this topic.