Hi All,
Is there a way to reload or restart the kernel without having to reboot?
Many thanks, -T
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
-T
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
On 9/26/20 5:51 PM, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
Hope you can provide any references :) It sounds interesting.
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
On Sat, 2020-09-26 at 01:51 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
There is a long-standing project that aims to do something like that ( https://github.com/dynup/kpatch), but AFAIK it's not production quality so far. Personally, I'm sceptical that it will ever be useful except in very constrained conditions. For one thing, it's not clear that there's much demand for it.
poc
在 2020-09-26星期六的 00:38 -0700，ToddAndMargo via users写道：
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
You can reload the whole OS by kexec, which is not considered as reboot (technically) since you don't redo the whole booting process and directly go to new kernel. (and all userspace process are killed, just like a normal reboot)
But I am pretty sure that this isn't the answer you are looking for, if you are finding a way to replace/change the current kernel without affecting userland, yes, but it can be tricky. We have kpatch and similar things, they are not that powerful to handle all kernel updates (those live patch tools are mostly for security patches, and if the update add/removed a kernel func, changed a internel structure, they may not work). They can override methods and can't do further things, yet.
Many thanks, -T
--
When we ask for advice, we are usually looking for an accomplice. -- Charles Varlet de La Grange
users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
ToddAndMargo via users writes:
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
Highly unlikely that this will ever happen, for a whole variety of technical reasons. This is pretty much a technical impossibility.
Patrick O'Callaghan writes:
On Sat, 2020-09-26 at 01:51 -0700, ToddAndMargo via users wrote:
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
There is a long-standing project that aims to do something like that ( https://github.com/dynup/kpatch), but AFAIK it's not production quality so far. Personally, I'm sceptical that it will ever be useful except in very constrained conditions. For one thing, it's not clear that there's much demand for it.
No, that's not what this is trying to do. Not even close.
This project attempts to implement the ability to patch the running kernel, in a number of limited situations.
This is nowhere close to loading a brand new kernel and somehow seamlessly switching to it.
On Sat, 2020-09-26 at 11:45 -0400, Sam Varshavchik wrote:
Patrick O'Callaghan writes:
On Sat, 2020-09-26 at 01:51 -0700, ToddAndMargo via users wrote:
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting. That ever happen?
There is a long-standing project that aims to do something like that ( https://github.com/dynup/kpatch), but AFAIK it's not production quality so far. Personally, I'm sceptical that it will ever be useful except in very constrained conditions. For one thing, it's not clear that there's much demand for it.
No, that's not what this is trying to do. Not even close.
This project attempts to implement the ability to patch the running kernel, in a number of limited situations.
This is nowhere close to loading a brand new kernel and somehow seamlessly switching to it.
Fair enough. I was never that interested so didn't take the trouble to read the details.
poc
On Sat, 2020-09-26 at 11:45 -0400, Sam Varshavchik wrote:
Patrick O'Callaghan writes:
There is a long-standing project that aims to do something like that ( https://github.com/dynup/kpatch), but AFAIK it's not production quality so far. Personally, I'm sceptical that it will ever be useful except in very constrained conditions. For one thing, it's not clear that there's much demand for it.
No, that's not what this is trying to do. Not even close.
This project attempts to implement the ability to patch the running kernel, in a number of limited situations.
This is nowhere close to loading a brand new kernel and somehow seamlessly switching to it.
For completeness, I'll mention that there are solutions that semantically reboot into a brand new kernel but transfer the userspace state to it with a minimum of actual downtime:
https://criu.org/Seamless_kernel_upgrade
This is like hibernating and immediately resuming except that the userspace state is transferred in a format that is compatible across kernel versions. However, it's unclear if any of these solutions are reasonable for typical end users.
Matt
On 2020-09-26 03:20, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
Usually that slow down is swapping. Have you checked to see how much swap is in use?
I have machines with 10gb of ram and that is not enough and used to get slowdowns prior to me setting up earlyoom. With earlyoom it generally kills sets of firefox tabs and generally works ok, but I have had things get bad enough that it keeps killing the tab I want to view and then I have to find someone else to kill so I have enough ram.
if you have sysstat / sar installed it would have data you could view with sar -S that would tell you if it is swapping/paging.
vmstat 1 swpd column, that is the numberof kb swapped, the higher it gets the uglier it gets.
And if it is the most likely memory usage, rebooting the kernel would not fix anything as memory consumption is mostly a userspace problem.
On Sat, Sep 26, 2020 at 6:12 PM ToddAndMargo via users users@lists.fedoraproject.org wrote:
On 2020-09-26 03:20, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue. _______________________________________________ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
On 2020-09-26 03:20, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
On 2020-09-26 16:22, Roger Heflin wrote:
Usually that slow down is swapping. Have you checked to see how much swap is in use?
top - 17:39:41 up 1:55, 1 user, load average: 0.26, 0.37, 0.37
Tasks: 245 total, 2 running, 243 sleeping, 0 stopped, 0 zombie
%Cpu(s): 3.6 us, 0.8 sy, 0.0 ni, 95.2 id, 0.2 wa, 0.2 hi, 0.1 si, 0.0 st
MiB Mem : 15896.7 total, 11478.9 free, 1867.8 used, 2550.0 buff/cache
MiB Swap: 8031.0 total, 8031.0 free, 0.0 used. 13333.0 avail Mem
I have machines with 10gb of ram and that is not enough and used to get slowdowns prior to me setting up earlyoom. With earlyoom it generally kills sets of firefox tabs and generally works ok, but I have had things get bad enough that it keeps killing the tab I want to view and then I have to find someone else to kill so I have enough ram.
When it happens, it is instant
if you have sysstat / sar installed it would have data you could view with sar -S that would tell you if it is swapping/paging.
vmstat 1 swpd column, that is the numberof kb swapped, the higher it gets the uglier it gets.
$ vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu----- r b swpd free buff cache si so bi bo in cs us sy id wa st 1 0 0 11738732 150112 2474552 0 0 70 21 289 516 3 2 95 0 0 0 0 0 11749568 150112 2462464 0 0 0 0 615 1458 1 1 97 0 0 0 0 0 11749708 150112 2462464 0 0 0 0 488 1016 1 1 98 0 0 1 0 0 11749820 150128 2462464 0 0 0 296 478 1185 1 1 98 0 0 0 0 0 11749820 150128 2462480 0 0 0 0 477 1068 2 1 98 0 0 0 0 0 11749568 150128 2462464 0 0 0 8 480 1043 1 1 99 0 0 0 0 0 11749048 150144 2462464 0 0 0 244 693 1425 2 1 97 0 0
And if it is the most likely memory usage, rebooting the kernel would not fix anything as memory consumption is mostly a userspace problem.
Also, the mouse chasing the pop up menues has nothing to do with the file system.
I am at a loss. If it was a hardware issue, a reboot would not fix it.
And my almost identical shop computer does not do this.
On Sat, 2020-09-26 at 16:11 -0700, ToddAndMargo via users wrote:
On 2020-09-26 03:20, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote:
Hi All,
Is there a way to reload or restart the kernel without having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
It's highly likely that the issue is not being fixed by the kernel rebooting as by processes being killed and restarted, which is what happens when you reboot. A live reboot that doesn't restart processes is probably not going to help.
poc
On Sat, 2020-09-26 at 17:44 -0700, ToddAndMargo via users wrote:
And if it is the most likely memory usage, rebooting the kernel would
not fix anything as memory consumption is mostly a userspace problem.
Also, the mouse chasing the pop up menues has nothing
to do with the file system.
I am at a loss. If it was a hardware issue, a reboot would not fix it.
There are several things you can test to try to narrow it down: - Log out and in again - Restart the DE without rebooting - Try a different DE
poc
On 2020-09-27 03:29, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 17:44 -0700, ToddAndMargo via users wrote:
And if it is the most likely memory usage, rebooting the kernel would
not fix anything as memory consumption is mostly a userspace problem.
Also, the mouse chasing the pop up menues has nothing
to do with the file system.
I am at a loss. If it was a hardware issue, a reboot would not fix it.
There are several things you can test to try to narrow it down:
- Log out and in again
- Restart the DE without rebooting
- Try a different DE
poc
Have not tried the log out and back in thing.
What is a "DE"
From the vmstat, all of the time is idle, you do not have any wait, generally if you have a disk having issues the iowait will go up, but you have zero.
Also from vmstat, the system time is low.
So the gui gets slow, a reboot clears it, does it clear up?
You might want to include a messages file from +=5 minutes of the the issue happening.
Weird kernel bugs causing kernel lock issues would produce this behavior, and would don't usually show on vmstat, but they can be seen in the messages file.
On Sat, Sep 26, 2020 at 7:51 PM ToddAndMargo via users users@lists.fedoraproject.org wrote:
On 2020-09-26 03:20, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 01:45 -0700, ToddAndMargo via users wrote:
On 2020-09-26 01:41, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 00:38 -0700, ToddAndMargo via users wrote: > Hi All, > > Is there a way to reload or restart the kernel without > having to reboot?
No.
poc
Poop!
Thank you for the confirmation.
Is the file system part of the kernel? If not, is there a way to restart it without rebooting?
Depends what you mean by "the file system". Some filesystems are reloadable (either as a module or as FUSE), but others aren't. The underlying filesystem layer is a fixed part of the kernel.
Maybe if you explained what you're trying to do, someone might be able to help.
poc
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
On 2020-09-26 16:22, Roger Heflin wrote:
Usually that slow down is swapping. Have you checked to see how much swap is in use?
top - 17:39:41 up 1:55, 1 user, load average: 0.26, 0.37, 0.37
Tasks: 245 total, 2 running, 243 sleeping, 0 stopped, 0 zombie
%Cpu(s): 3.6 us, 0.8 sy, 0.0 ni, 95.2 id, 0.2 wa, 0.2 hi, 0.1 si, 0.0 st
MiB Mem : 15896.7 total, 11478.9 free, 1867.8 used, 2550.0 buff/cache
MiB Swap: 8031.0 total, 8031.0 free, 0.0 used. 13333.0 avail Mem
I have machines with 10gb of ram and that is not enough and used to get slowdowns prior to me setting up earlyoom. With earlyoom it generally kills sets of firefox tabs and generally works ok, but I have had things get bad enough that it keeps killing the tab I want to view and then I have to find someone else to kill so I have enough ram.
When it happens, it is instant
if you have sysstat / sar installed it would have data you could view with sar -S that would tell you if it is swapping/paging.
vmstat 1 swpd column, that is the numberof kb swapped, the higher it gets the uglier it gets.
$ vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu----- r b swpd free buff cache si so bi bo in cs us sy id wa st 1 0 0 11738732 150112 2474552 0 0 70 21 289 516 3 2 95 0 0 0 0 0 11749568 150112 2462464 0 0 0 0 615 1458 1 1 97 0 0 0 0 0 11749708 150112 2462464 0 0 0 0 488 1016 1 1 98 0 0 1 0 0 11749820 150128 2462464 0 0 0 296 478 1185 1 1 98 0 0 0 0 0 11749820 150128 2462480 0 0 0 0 477 1068 2 1 98 0 0 0 0 0 11749568 150128 2462464 0 0 0 8 480 1043 1 1 99 0 0 0 0 0 11749048 150144 2462464 0 0 0 244 693 1425 2 1 97 0 0
And if it is the most likely memory usage, rebooting the kernel would not fix anything as memory consumption is mostly a userspace problem.
Also, the mouse chasing the pop up menues has nothing to do with the file system.
I am at a loss. If it was a hardware issue, a reboot would not fix it.
And my almost identical shop computer does not do this.
users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
On Sun, 2020-09-27 at 04:26 -0700, ToddAndMargo via users wrote:
On 2020-09-27 03:29, Patrick O'Callaghan wrote:
On Sat, 2020-09-26 at 17:44 -0700, ToddAndMargo via users wrote:
And if it is the most likely memory usage, rebooting the kernel would
not fix anything as memory consumption is mostly a userspace problem.
Also, the mouse chasing the pop up menues has nothing
to do with the file system.
I am at a loss. If it was a hardware issue, a reboot would not fix it.
There are several things you can test to try to narrow it down:
- Log out and in again
- Restart the DE without rebooting
- Try a different DE
poc
Have not tried the log out and back in thing.
What is a "DE"
Desktop Environment, i.e. the GUI and associated stuff you use to interact with the system. You said you were using Xfce. You could try KDE or Gnome among others as a test. If they're installed, it's just a matter of selecting one from the login screen. Your other software should continue to work.
poc
ToddAndMargo via users writes:
A reboot ALWAYS fixes the issue.
That's probably because it reinitializes *all* the data structures, in the kernel, in system processes, and in user processes. Evidently clearing state is what fixes your problem.
I am trying to get around the reboot thing.
Loading the kernel is very fast. What takes time is initializing hardware and all the data structures. Since you don't know where the problem is, that's really the only sure way.
Somebody (quite possibly you, since you're the one with direct access to the machine) needs to find out what aspect of state (probably not swap from your data) is causing the slowdown. Have you checked for bitcoin miners? ;-)
On 2020-09-27 21:23, Stephen J. Turnbull wrote:
Somebody (quite possibly you, since you're the one with direct access to the machine) needs to find out what aspect of state (probably not swap from your data) is causing the slowdown
How do I do that?
Also, I ran
# lynis audit system and # /usr/bin/chkrootkit
No malware located
On Sun, 2020-09-27 at 21:40 -0700, ToddAndMargo via users wrote:
although Firefox is slow to open tabs, by what is new about that?
How many do you have open?
On Sat, Sep 26, 2020 at 04:11:33PM -0700, ToddAndMargo via users wrote:
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
Brendan Gregg has written several good talks and documents about various Linux performance measurement tools. You might want to see what the kernel is doing when you see sluggish behavior.
http://brendangregg.com/linuxperf.html
It does require some deep spelunking into the kernel internals, but it is actually quite amazing what the kernel has for monitoring its activities. I use it quite often to debug filesystem behavior, just poke around /sys/kernel/debug/tracing/events/.
https://www.kernel.org/doc/html/v5.8/trace/events.html
From my experience, as an end user providing feedback, the kernel developers really like it when you can point out a particular syscall from the trace output that is misbehaving.
On Sat, Sep 26, 2020 at 01:51:59AM -0700, ToddAndMargo via users wrote:
Somewhere I read that they were working on the ability to start a new kernel after a dnf upgrde without rebooting.
The reason we want to do this is to avoid the need to re-enter your disk encryption passphrase. It'll still mean everything restarts in this case.
On 2020-09-28 05:49, Jonathan Billings wrote:
On Sat, Sep 26, 2020 at 04:11:33PM -0700, ToddAndMargo via users wrote:
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
Brendan Gregg has written several good talks and documents about various Linux performance measurement tools. You might want to see what the kernel is doing when you see sluggish behavior.
http://brendangregg.com/linuxperf.html
It does require some deep spelunking into the kernel internals, but it is actually quite amazing what the kernel has for monitoring its activities. I use it quite often to debug filesystem behavior, just poke around /sys/kernel/debug/tracing/events/.
Anything is particular to look at?
https://www.kernel.org/doc/html/v5.8/trace/events.html
From my experience, as an end user providing feedback, the kernel developers really like it when you can point out a particular syscall from the trace output that is misbehaving.
On Tue, 29 Sep 2020 at 09:26, ToddAndMargo via users < users@lists.fedoraproject.org> wrote:
On 2020-09-28 05:49, Jonathan Billings wrote:
On Sat, Sep 26, 2020 at 04:11:33PM -0700, ToddAndMargo via users wrote:
FC 32, x64 Ext4 Xfce 4.14
Occasionally, my computers slows down. I have not been able to pin down why. Top shows very little memory usage.
A 1 hour, 12 minute dump (dump/restore) takes 15 hours when this happens. Pop up menus start to lag behind the mouse
I am trying to get around the reboot thing.
-T
A reboot ALWAYS fixes the issue.
Brendan Gregg has written several good talks and documents about various Linux performance measurement tools. You might want to see what the kernel is doing when you see sluggish behavior.
This is an excellent site. For starters, though, bpytop gives an overview that may help focus your investigation.
It does require some deep spelunking into the kernel internals, but it is actually quite amazing what the kernel has for monitoring its activities. I use it quite often to debug filesystem behavior, just poke around /sys/kernel/debug/tracing/events/.
Anything is particular to look at?
For starters: temperature spikes (numerous utilities, including bpytop) and disk errors (smartmontools). This doesn't appear to be a widespread problem, so you should think about malware, uncommon hardware, uncommon software.
If you have the resources to swap out the storage and start with a fresh install to see if the problem goes away that would help narrow the search.
https://www.kernel.org/doc/html/v5.8/trace/events.html
From my experience, as an end user providing feedback, the kernel developers really like it when you can point out a particular syscall from the trace output that is misbehaving.
Very true.
-
George N. White III -
Jonathan Billings -
Matt McCutchen -
Matthew Miller -
nico＠niconi.org -
Patrick O'Callaghan -
Roger Heflin -
Sam Varshavchik -
Stephen J. Turnbull -
Tim -
ToddAndMargo -
WooHyung Jeon