I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit IO scheduler if cfq The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable. I have monitors for CPU and network and disk activity, and I've tried this when all of them are low. I've tried this with each CPU activity showing ... Again low, single digits or idle cores. Within the KDE desktop, switching between desktops is fast, moving though FF tabs is fast, moving though the (presumably cached) messages and folders in T'Bird is fast. Perhaps I shouldn't complain. But it strikes me as odd. I'm wondering if this is related to the ssh...Login/logind issue in another thread. However the latency is there even if I'm switching to vt1 and have a already logged in there or swiching back to the X/KDE on vt7. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable. This is a problem with the schedule or counsel-kit? Is the schedule the one in
On 03/12/2015 11:47 AM, Anton Aylward wrote: the kernel of is this the Pottering scripts that objected to the CFS -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 12:02 PM, Ruben Safir wrote:
On 03/12/2015 11:47 AM, Anton Aylward wrote:
The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable.
This is a problem with the schedule or counsel-kit?
I presume you mean "scheduler or console-kit". Good question. I don't know and can't think of a way that I can tell.
Is the schedule the one in the kernel of is this the Pottering scripts that objected to the CFS
Sorry, I don't even understand that question. I've tried this with other settings - none and NOOP and Deadline - and the latency is still there. What scripts are you referring to and how might I be running them? If you are referring to this http://en.wikipedia.org/wiki/Completely_Fair_Scheduler#Controversy then the answer is a 'no, not that I know of, not by an active decision I've taken to do so'. All I'm doing for this simple test is hotkeying between vts My observation is that it didn't used to be like this .... Last year. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
All I'm doing for this simple test is hotkeying between vts My observation is that it didn't used to be like this .... Last year. Yeah that behavior has been there since systemd took over the getty's attachments to the counsels. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 02:41 PM, Ruben wrote:
All I'm doing for this simple test is hotkeying between vts My observation is that it didn't used to be like this .... Last year.
Yeah that behavior has been there since systemd took over the getty's attachments to the counsels.
NO! Last year I was using systemd. Its come about because of a kernel update... -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward composed on 2015-03-12 14:59 (UTC-0400):
Ruben wrote:
All I'm doing for this simple test is hotkeying between vts My observation is that it didn't used to be like this .... Last year.
Yeah that behavior has been there since systemd took over the getty's attachments to the counsels.
NO!
Last year I was using systemd. Its come about because of a kernel update...
If you're more interested in making problem go away than finding out why the problem, you could do what I've been doing for over a year, switching from on-demand gettys to always ready gettys: # /bin/bash cd /etc/systemd/system/getty.target.wants cp -a ../../../../usr/lib/systemd/system/getty@.service ../../../../usr/lib/systemd/system/getty@tty1.service sed -i 's/TTYVTDisallocate=yes/TTYVTDisallocate=no/' ../../../../usr/lib/systemd/system/getty@tty1.service ln -sf ../../../../usr/lib/systemd/system/getty@tty1.service ./getty@tty1.service systemctl start getty@tty1.service ln -sf ../../../../usr/lib/systemd/system/getty@.service ./getty@tty2.service systemctl start getty@tty2.service ln -sf ../../../../usr/lib/systemd/system/getty@.service ./getty@tty3.service systemctl start getty@tty3.service ln -sf ../../../../usr/lib/systemd/system/getty@.service ./getty@tty4.service systemctl start getty@tty4.service ln -sf ../../../../usr/lib/systemd/system/getty@.service ./getty@tty5.service systemctl start getty@tty5.service ln -sf ../../../../usr/lib/systemd/system/getty@.service ./getty@tty6.service systemctl start getty@tty6.service Not likely it would have any impact on ttys already logged in on, but who knows. I do think latency issues generally have grown more widespread since a release or three back, maybe around the time the last vestiges of sysvinit began being purged. Are there any clues on tty10 or in dmesg? Which kernel were you using before 3.19.x? Was it OK? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 03:21 PM, Felix Miata wrote:
If you're more interested in making problem go away than finding out why the problem, you could do what I've been doing for over a year, switching from on-demand gettys to always ready gettys:
# /bin/bash cd /etc/systemd/system/getty.target.wants cp -a ../../../../usr/lib/systemd/system/getty@.service ../../../../usr/lib/systemd/system/getty@tty1.service sed -i 's/TTYVTDisallocate=yes/TTYVTDisallocate=no/' ../../../../usr/lib/systemd/system/getty@tty1.service ln -sf ../../../../usr/lib/systemd/system/getty@tty1.service ./getty@tty1.service systemctl start getty@tty1.service
Er-um.. I have that. That's what I've beet talking about lrwxrwxrwx 1 root root 38 Nov 6 2013 /etc/systemd/system/getty.target.wants/getty@tty1.service -> /usr/lib/systemd/system/getty@.service # systemctl status getty@tty1.service getty@tty1.service - Getty on tty1 Loaded: loaded (/usr/lib/systemd/system/getty@.service; enabled) Drop-In: /usr/lib/systemd/system/getty@tty1.service.d âânoclear.conf Active: active (running) since Mon 2015-03-09 09:53:23 EDT; 3 days ago Docs: man:agetty(8) man:systemd-getty-generator(8) http://0pointer.de/blog/projects/serial-console.html Main PID: 1692 (login) CGroup: /system.slice/system-getty.slice/getty@tty1.service ⣠1692 login -- root Similar for tty2 and so on. Switching between 1&2 or 1&12 (where the log goes to, not 10) is rapid by comparison.
I do think latency issues generally have grown more widespread since a release or three back, maybe around the time the last vestiges of sysvinit began being purged.
Within 13.1? -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward composed on 2015-03-12 17:43 (UTC-0400):
I do think latency issues generally have grown more widespread since a release or three back, maybe around the time the last vestiges of sysvinit began being purged.
Within 13.1?
Don't remember. If the timestamp on my oldest saved version of that gettys script is any indication, could be, as it was 20 months ago. Might be I started futzing with it in Cauldron or Rawhide, as openSUSE's historically been throttling implementation of systemd evolution. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
IO scheduler if cfq
The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable.
Yikes, where are you getting your kernels? I'm on OS 13.2 64 bit, and using kernel 3.16.7-7-desktop. Switching is instantaneous. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 01:03 PM, John Andersen wrote:
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
IO scheduler if cfq
The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable.
Yikes, where are you getting your kernels? I'm on OS 13.2 64 bit, and using kernel 3.16.7-7-desktop.
http://download.opensuse.org/repositories/Kernel:/openSUSE-13.1/standard/
Switching is instantaneous.
Well, given that its not any of the scheduler option, since I've tried them all, and assuming its not something else in the delta between 13.1 and 13.2, then it looks like my issue of the kernel being responsible is a reasonable assumption. I'm going to to get hold of 3.16 series and see. .... ..... ...... Well, it seems the current repositories for 13.1 just pull the latest GIT or have the 3.11 that was on the baseline distribution. It looks like my weekend project, if I chose to accept it, is looking though the change logs to see what's in the 3.19 that might account for this. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Thu, Mar 12, 2015 at 2:57 PM, Anton Aylward <opensuse@antonaylward.com> wrote:
On 03/12/2015 01:03 PM, John Andersen wrote:
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
IO scheduler if cfq
The last few kernel revisions seem to have affected latency in certain areas. This is most obvious in hot keying from the GUI screen on vt7 to a plain text mode login on vt1. I should be almost instantaneous. I was in past months. Its now taking 3-4 seconds. That really really noticeable.
Yikes, where are you getting your kernels? I'm on OS 13.2 64 bit, and using kernel 3.16.7-7-desktop.
http://download.opensuse.org/repositories/Kernel:/openSUSE-13.1/standard/
Switching is instantaneous.
Well, given that its not any of the scheduler option, since I've tried them all, and assuming its not something else in the delta between 13.1 and 13.2, then it looks like my issue of the kernel being responsible is a reasonable assumption.
I'm going to to get hold of 3.16 series and see. .... ..... ...... Well, it seems the current repositories for 13.1 just pull the latest GIT or have the 3.11 that was on the baseline distribution.
It looks like my weekend project, if I chose to accept it, is looking though the change logs to see what's in the 3.19 that might account for this.
Sometimes newer kernels require userspace app changes to go with them. I doubt if many people are testing 3.19 kernels with oS 13.1 But you are not the only one: http://markmail.org/thread/y6jgpfkclc6akpxy Note that in the above the 3.19 kernel seems to have broken all USB external storage for now. Bruno looks like he is in the process of fixing that. http://markmail.org/message/23rfrmyg45xaydrq Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Thursday 12 Mar 2015 16:08:51 Greg Freemyer wrote:
On Thu, Mar 12, 2015 at 2:57 PM, Anton Aylward
<opensuse@antonaylward.com> wrote:
On 03/12/2015 01:03 PM, John Andersen wrote:
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
snip Note that in the above the 3.19 kernel seems to have broken all USB external storage for now. Bruno looks like he is in the process of fixing that.
Is this USB issue only realated to Storage? i've got a bug 920937 lodged against the system not being able to recognise a USB printer connect on the USB.
Greg
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, Mar 13, 2015 at 7:38 AM, ianseeks <ianseeks@yahoo.co.uk> wrote:
On Thursday 12 Mar 2015 16:08:51 Greg Freemyer wrote:
On Thu, Mar 12, 2015 at 2:57 PM, Anton Aylward
<opensuse@antonaylward.com> wrote:
On 03/12/2015 01:03 PM, John Andersen wrote:
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
snip Note that in the above the 3.19 kernel seems to have broken all USB external storage for now. Bruno looks like he is in the process of fixing that.
Is this USB issue only realated to Storage? i've got a bug 920937 lodged against the system not being able to recognise a USB printer connect on the USB.
The problem is specific to running the bright new shiny 3.19 kernel on the nice stable well worn openSUSE 13.1 release. If that is what you are trying to do, I'd say it is buyer beware..... Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Friday 13 Mar 2015 12:35:00 Greg Freemyer wrote:
On Fri, Mar 13, 2015 at 7:38 AM, ianseeks <ianseeks@yahoo.co.uk> wrote:
On Thursday 12 Mar 2015 16:08:51 Greg Freemyer wrote:
On Thu, Mar 12, 2015 at 2:57 PM, Anton Aylward
<opensuse@antonaylward.com> wrote:
On 03/12/2015 01:03 PM, John Andersen wrote:
On 03/12/2015 08:47 AM, Anton Aylward wrote:
I'm currently running 3.19.1-1-desktop #1 SMP PREEMPT on a Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz With 13.1/64-bit
snip
Note that in the above the 3.19 kernel seems to have broken all USB external storage for now. Bruno looks like he is in the process of fixing that.
Is this USB issue only realated to Storage? i've got a bug 920937 lodged against the system not being able to recognise a USB printer connect on the USB.
The problem is specific to running the bright new shiny 3.19 kernel on the nice stable well worn openSUSE 13.1 release.
If that is what you are trying to do, I'd say it is buyer beware.....
i'm using Tumbleweed with 3.19, but when it was 3.18 it worked with the printer connected via USB
Greg
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 04:08 PM, Greg Freemyer wrote:
I doubt if many people are testing 3.19 kernels with oS 13.1
But you are not the only one: http://markmail.org/thread/y6jgpfkclc6akpxy
Note that in the above the 3.19 kernel seems to have broken all USB external storage for now. Bruno looks like he is in the process of fixing that.
I don't see any of this. I've been plugging in old (256M) USP sticks and modern 32G ones just this last week. I've formatted them both as vfat and ext3 within the last month. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward wrote:
It looks like my weekend project, if I chose to accept it, is looking though the change logs to see what's in the 3.19 that might account for this.
Ditto here. i'm running a 3.18.5 right now, but have noticed some odd latency in simply things, like 'cd' into a directory, or 'ls' of a small dir. Taking ~2-3 seconds in some cases where it should be instantaneous. I'm not running systemd.... It's unfortunate 3.19 shows a similar problem as I'd just made a 3.19.1 kernel -- just haven't switched to it yet. Tried running xlatencytop to see if I could catch anything but...no luck -- it really needs to log... I thought I'd done something horrid to my system. But now... very interesting! Thanks for the post! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 03/12/2015 07:56 PM, Linda Walsh wrote:
I'm not running systemd.... It's unfortunate 3.19 shows how are you doing that and staying up to date? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Ruben wrote:
On 03/12/2015 07:56 PM, Linda Walsh wrote:
I'm not running systemd.... It's unfortunate 3.19 shows how are you doing that and staying up to date?
Same way I have been updating Suse since my 12.1 debacle. Manually... packages at a time. Inspect the ones that I suspect might mess w/things, Have a few files set to 'immutable', so if something tries to change them, they likely "won't". Have a few backup copies of {,/usr}/{lib{,64},{s,}bin}. Don't have most of 13.2 installed yet.. but 13.1 wasn't horrible. It's not perfect, but life goes on. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (8)
-
Anton Aylward -
Felix Miata -
Greg Freemyer -
ianseeks -
John Andersen -
Linda Walsh -
Ruben -
Ruben Safir