When I run xenmon.py on my x86_64 box (dual core, single socket), I get a single line of incomprehensible output at the top of the screen. It looks something like this (extra spaces & =''s removed to make it fit): = 99.96% 99.98%======18.70ms/exx======Waitedd=========== Is this a known problem? What is the output supposed to look like? -- Ray Bryant AMD Performance Labs Austin, Tx 512-602-0038 (o) 512-507-7807 (c) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
I haven''t seen that behavior. In 3.0.2, I had to change the xenbaked
call in
xenmon so that it would redirect stdout to /dev/null as it kept outputting
some debugging code that I presume accidently was left in.
The output should look something like this (note, I''m only display the
left
half of the screen as otherwise it would wrap in e-mail and look awful based
on my terminal width):
CPU = 0 Last 10 seconds
=========================================== 0 775.49 ns 0.00% 272.49
us/ex
0 6.24 ms/ex
0 59.38 us 0.01% 0.00 ns/io
0 13.87 ns 0.00% 4.87 us/ex
0 0/s
0 0/s 0/ex
31 59.39 us 0.01% 20.87 ms/ex
31 992.84 ms/ex
31 130.57 ns 0.00% 0.00 ns/io
31 777.88 ns 0.00% 273.33 us/ex
31 0/s
31 0/s 0/ex
0.01%
I''m in the middle of playing with all the monitoring capabilities and
am
writing up a tutorial, of sorts, on this functionality.
There are a few things that you might find relevant in the following:
http://kibab.homeip.net/hw/vienna_hw6/tutorial.html
It''s a work in progress for one of my classes, but is hopefully decent.
--Kaleb
On Wednesday 10 May 2006 1:20 pm, Ray Bryant wrote:> When I run xenmon.py on my x86_64 box (dual core, single socket), I get a
> single line of incomprehensible output at the top of the screen. It
> looks something like this (extra spaces & =''s removed to make
it fit):
>
> = 99.96% 99.98%======18.70ms/exx======Waitedd===========>
> Is this a known problem? What is the output supposed to look like?
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
And what could be the reason fort this: #xenmon.py xenbaked: no process killed ms_per_sample = 100 Initialized with 2 cpu''s CPU Frequency = 2388.52 ERROR: Failure to get trace buffer pointer from Xen (22 = Invalid argument) xenbaked: no process killed ==========================0.00%========== Thanks. On 5/10/06, Kaleb Pederson <kibab@icehouse.net> wrote:> I haven''t seen that behavior. In 3.0.2, I had to change the xenbaked call in > xenmon so that it would redirect stdout to /dev/null as it kept outputting > some debugging code that I presume accidently was left in. > > The output should look something like this (note, I''m only display the left > half of the screen as otherwise it would wrap in e-mail and look awful based > on my terminal width): > > CPU = 0 Last 10 seconds > ===========================================> 0 775.49 ns 0.00% 272.49 us/ex > 0 6.24 ms/ex > 0 59.38 us 0.01% 0.00 ns/io > 0 13.87 ns 0.00% 4.87 us/ex > 0 0/s > 0 0/s 0/ex > 31 59.39 us 0.01% 20.87 ms/ex > 31 992.84 ms/ex > 31 130.57 ns 0.00% 0.00 ns/io > 31 777.88 ns 0.00% 273.33 us/ex > 31 0/s > 31 0/s 0/ex > 0.01% > > I''m in the middle of playing with all the monitoring capabilities and am > writing up a tutorial, of sorts, on this functionality. > > There are a few things that you might find relevant in the following: > > http://kibab.homeip.net/hw/vienna_hw6/tutorial.html > > It''s a work in progress for one of my classes, but is hopefully decent. > > --Kaleb > > On Wednesday 10 May 2006 1:20 pm, Ray Bryant wrote: > > When I run xenmon.py on my x86_64 box (dual core, single socket), I get a > > single line of incomprehensible output at the top of the screen. It > > looks something like this (extra spaces & =''s removed to make it fit): > > > > = 99.96% 99.98%======18.70ms/exx======Waitedd===========> > > > Is this a known problem? What is the output supposed to look like? > > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users > > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Serge, Sounds like you haven''t set your xen trace stuff running (not sure of the technical term...) Try running these commands: # tbctl 1 # setsize 10 then run xenmon and it should work. Kaleb - your tutorial looks interesting, I look forward to reading more of it when it is complete. On 5/11/06, Serge Dubrouski <sergeyfd@gmail.com> wrote:> > And what could be the reason fort this: > > #xenmon.py > xenbaked: no process killed > ms_per_sample = 100 > Initialized with 2 cpu''s > CPU Frequency = 2388.52 > ERROR: Failure to get trace buffer pointer from Xen (22 = Invalid > argument) > xenbaked: no process killed > > > ==========================0.00%==========> > Thanks. > > On 5/10/06, Kaleb Pederson <kibab@icehouse.net> wrote: > > I haven''t seen that behavior. In 3.0.2, I had to change the xenbaked > call in > > xenmon so that it would redirect stdout to /dev/null as it kept > outputting > > some debugging code that I presume accidently was left in. > > > > The output should look something like this (note, I''m only display the > left > > half of the screen as otherwise it would wrap in e-mail and look awful > based > > on my terminal width): > > > > CPU = 0 Last 10 seconds > > ===========================================> > 0 775.49 ns 0.00% 272.49 us/ex > > 0 6.24 ms/ex > > 0 59.38 us 0.01% 0.00 ns/io > > 0 13.87 ns 0.00% 4.87 us/ex > > 0 0/s > > 0 0/s 0/ex > > 31 59.39 us 0.01% 20.87 ms/ex > > 31 992.84 ms/ex > > 31 130.57 ns 0.00% 0.00 ns/io > > 31 777.88 ns 0.00% 273.33 us/ex > > 31 0/s > > 31 0/s 0/ex > > 0.01% > > > > I''m in the middle of playing with all the monitoring capabilities and am > > writing up a tutorial, of sorts, on this functionality. > > > > There are a few things that you might find relevant in the following: > > > > http://kibab.homeip.net/hw/vienna_hw6/tutorial.html > > > > It''s a work in progress for one of my classes, but is hopefully decent. > > > > --Kaleb > > > > On Wednesday 10 May 2006 1:20 pm, Ray Bryant wrote: > > > When I run xenmon.py on my x86_64 box (dual core, single socket), I > get a > > > single line of incomprehensible output at the top of the screen. It > > > looks something like this (extra spaces & =''s removed to make it fit): > > > > > > = 99.96% 99.98%======18.70ms/exx======Waitedd===========> > > > > > Is this a known problem? What is the output supposed to look like? > > > > > > _______________________________________________ > > Xen-users mailing list > > Xen-users@lists.xensource.com > > http://lists.xensource.com/xen-users > > > > > > > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Thursday 11 May 2006 10:48, Tim Wood wrote:> Serge, > Sounds like you haven''t set your xen trace stuff running (not sure of the > technical term...) > Try running these commands: > > # tbctl 1 > # setsize 10 > > then run xenmon and it should work.Actually, it appears that now one must enable the tracing on the xen command in grub/menu.lst as follows: kernel /xen-3.0-unstable.gz tbuf_size=10 I don''t when this change went in, but it appears to be required for the current xenbits unstable tree. (Look for the string "tbuf_size" in common/trace.c in the xen tree...) -- Ray Bryant AMD Performance Labs Austin, Tx 512-602-0038 (o) 512-507-7807 (c) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users