Re: I/O Bound Procs Clobbering Responsiveness

2011-12-01 Thread Patrick Lists
On 01-12-11 15:57, Ankur Sinha wrote:
> hello all,
>
> Quite a few people have observed screen freezes when there is heavy
> i/o.. This is the related bug:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=721127
>
> Do have a look. There isn't a solution currently, they're still trying
> to diagnose the issue.

Thanks for the pointer. I have seen a number of screen freezes on my 
workstation, an Intel i7, 6GB ram, encrypted 2x 2TB SATA (6Gb speed) 
RAID1 (mdraid), ATI HD59xx with F16 x86_64.

Maybe I read a lot of pdf's but I have seen the majority of the freezes 
occur during resizing the Evince window or changing the size.

I was under the assumption that it was caused by the radeon driver 
because I noticed that the temp of my videocard went to nearly 80C 
during regular usage. What I saw was a total screen freeze including the 
mouse. The clock would also no longer change. And I got no responses to 
a ping or ssh from my laptop to the workstation. Can you please confirm 
that you have seen similar symptoms?

Because I thought that it was caused by the radeon driver I installed 
the proprietary Catalyst 11.11 drivers from rpmfusion. Since then I have 
no longer seen a freeze.

Regards,
Patrick
-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test

Re: I/O Bound Procs Clobbering Responsiveness

2011-12-01 Thread Ankur Sinha
On Tue, 2011-11-08 at 12:17 +0100, Karel Volný wrote:
> Hi,
> 
> I'm afraid the problem is not specific to Fedora ... I see 
> something similar on my Gentoo box after upgrading kernel from 
> 2.6.39 to 3.0.6
> 
> I haven't tried with USB but I can see whole system being pretty 
> slow if there is high SATA load.
> 
> And also I had to stop bitcoin computations on my GPU, as the GUI 
> became totally unresponisve while running poclbm - this didn't 
> happen with previous kernel, I could run the computations in the 
> background while working with the GUI.
> 
> /me is thinking about kernel downgrade ...
> 
> K.
> 
> On Monday 07 of November 2011 17:36:10 Chuck Forsberg WA7KGX 
> N2469R wrote:
> > This a subjective comment, but it seems that I/O bound procs
> > such as disk copy are seriously impacting GUI response more
> > than before.   For example, switching windows can take 30
> > seconds or more.  Indeed it appears the system has halted
> > except that the clock seconds keep ticking away.
> > 
> > Setup:  R6550 8GB Fedora 16 64 bit accessed via TigerVNC over
> > a gigabit LAN.
> 
> 

hello all,

Quite a few people have observed screen freezes when there is heavy
i/o.. This is the related bug:

https://bugzilla.redhat.com/show_bug.cgi?id=721127

Do have a look. There isn't a solution currently, they're still trying
to diagnose the issue. 
-- 
Thanks, 
Regards,
Ankur: "FranciscoD"

http://fedoraproject.org/wiki/User:Ankursinha
http://dodoincfedora.wordpress.com/



signature.asc
Description: This is a digitally signed message part
-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test

Re: I/O Bound Procs Clobbering Responsiveness

2011-11-08 Thread Karel Volný

Hi,

I'm afraid the problem is not specific to Fedora ... I see 
something similar on my Gentoo box after upgrading kernel from 
2.6.39 to 3.0.6

I haven't tried with USB but I can see whole system being pretty 
slow if there is high SATA load.

And also I had to stop bitcoin computations on my GPU, as the GUI 
became totally unresponisve while running poclbm - this didn't 
happen with previous kernel, I could run the computations in the 
background while working with the GUI.

/me is thinking about kernel downgrade ...

K.

On Monday 07 of November 2011 17:36:10 Chuck Forsberg WA7KGX 
N2469R wrote:
> This a subjective comment, but it seems that I/O bound procs
> such as disk copy are seriously impacting GUI response more
> than before.   For example, switching windows can take 30
> seconds or more.  Indeed it appears the system has halted
> except that the clock seconds keep ticking away.
> 
> Setup:  R6550 8GB Fedora 16 64 bit accessed via TigerVNC over
> a gigabit LAN.


-- 
Karel Volný
QE BaseOs/Daemons Team
Red Hat Czech, Brno
tel. +420 532294274
(RH: +420 532294111 ext. 8262074)
xmpp ka...@jabber.cz
:: "Never attribute to malice what can
::  easily be explained by stupidity."


signature.asc
Description: This is a digitally signed message part.
-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test

Re: I/O Bound Procs Clobbering Responsiveness

2011-11-08 Thread Peter Robinson
On Tue, Nov 8, 2011 at 7:08 AM, Jared K. Smith  wrote:
> On Mon, Nov 7, 2011 at 10:06 PM, Chuck Forsberg WA7KGX N2469R
>  wrote:
>> This a subjective comment, but it seems that I/O bound procs
>> such as disk copy are seriously impacting GUI response more
>> than before.
>
> I noticed the same thing yesterday while trying to create a live usb
> key using livecd-iso-to-disk, only it only seemed to be impacting
> certain GUIs.  While creating the key, firefox was completely
> unresponsive.  Gnome terminal seemed to work just fine.  Accessing the
> GNOME dash (either by pressing the meta key or moving the mouse to the
> hot corner) was unresponsive, but I could press CTRL-ALT-down to get
> to the second virtual desktop.
>
> In my quick tests in the airport this morning, the issue appears to be
> related to USB I/O (at least on my machine), as I can't reproduce the
> problem copying data from one SATA disk to another (using a SATA disk
> in the CD-ROM bay).  I'll try to do some more testing once I get back
> home.

I see the same thing with usb but I think I also see it when doing
disk to network transfers as well like for scp. My HDD is encrypted
which likely doesn't help perf but its not normally this bad. I've
been seeing it for a while, I can't say exactly when it started but
its possible or around when I upgraded from F-15 to F-16 beta.

Peter

Peter
-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test

Re: I/O Bound Procs Clobbering Responsiveness

2011-11-07 Thread Jared K. Smith
On Mon, Nov 7, 2011 at 10:06 PM, Chuck Forsberg WA7KGX N2469R
 wrote:
> This a subjective comment, but it seems that I/O bound procs
> such as disk copy are seriously impacting GUI response more
> than before.

I noticed the same thing yesterday while trying to create a live usb
key using livecd-iso-to-disk, only it only seemed to be impacting
certain GUIs.  While creating the key, firefox was completely
unresponsive.  Gnome terminal seemed to work just fine.  Accessing the
GNOME dash (either by pressing the meta key or moving the mouse to the
hot corner) was unresponsive, but I could press CTRL-ALT-down to get
to the second virtual desktop.

In my quick tests in the airport this morning, the issue appears to be
related to USB I/O (at least on my machine), as I can't reproduce the
problem copying data from one SATA disk to another (using a SATA disk
in the CD-ROM bay).  I'll try to do some more testing once I get back
home.

--
Jared Smith
Fedora Project Leader
-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test

I/O Bound Procs Clobbering Responsiveness

2011-11-07 Thread Chuck Forsberg WA7KGX N2469R
This a subjective comment, but it seems that I/O bound procs
such as disk copy are seriously impacting GUI response more
than before.   For example, switching windows can take 30 seconds
or more.  Indeed it appears the system has halted except that the
clock seconds keep ticking away.

Setup:  R6550 8GB Fedora 16 64 bit accessed via TigerVNC over
a gigabit LAN.

-- 
Chuck Forsberg WA7KGX N2469R c...@omen.com   www.omen.com
Developer of Industrial ZMODEM(Tm) for Embedded Applications
   Omen Technology Inc  "The High Reliability Software"
10255 NW Old Cornelius Pass Portland OR 97231   503-614-0430

-- 
test mailing list
test@lists.fedoraproject.org
To unsubscribe: 
https://admin.fedoraproject.org/mailman/listinfo/test