9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* [9fans] Fossil+venti install problem
@ 2008-01-09  8:17 a.verardo
  2008-01-09 13:11 ` erik quanstrom
  0 siblings, 1 reply; 10+ messages in thread
From: a.verardo @ 2008-01-09  8:17 UTC (permalink / raw)
  To: 9fans

On Dec 27 I wrote about a problem I had installing a
fossil+venti config using the ISO image downloaded on Dec
22:

* On several HW setup, from P500 to P3000, the
* install goes ok but just after the boot from HD
* starts an endless activity (> 20 h ?!) with a steady
* 100% cpu load and periodic (1 sec seek, 3 sec. pause)
* disk load. Then the machine hangs.

The same behaviour has also been observed in the University
lab by some students working on a thesis about Plan9.

Is there other 9fan(s) who had the same problem ?

a.v.


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09  8:17 [9fans] Fossil+venti install problem a.verardo
@ 2008-01-09 13:11 ` erik quanstrom
  0 siblings, 0 replies; 10+ messages in thread
From: erik quanstrom @ 2008-01-09 13:11 UTC (permalink / raw)
  To: 9fans

> On Dec 27 I wrote about a problem I had installing a
> fossil+venti config using the ISO image downloaded on Dec
> 22:
> 
> * On several HW setup, from P500 to P3000, the
> * install goes ok but just after the boot from HD
> * starts an endless activity (> 20 h ?!) with a steady
> * 100% cpu load and periodic (1 sec seek, 3 sec. pause)
> * disk load. Then the machine hangs.
> 
> The same behaviour has also been observed in the University
> lab by some students working on a thesis about Plan9.
> 
> Is there other 9fan(s) who had the same problem ?
> 
> a.v.

i hesitate to answer because i have not seen such a problem
nor do i run venti or fossil.

i haven't any answers, but perhaps these questions would help
narrow things down (assuming of course that you haven't hit an
actual venti bug):

1. what kind of drives are you using; what hba?  have you enabled dma?
2. which processes are accumulating cpu time?
3. what is the interrupt load, memory usage? ("stats -lsef")

- erik


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09 17:05 a.verardo
  2008-01-09 17:35 ` erik quanstrom
@ 2008-01-10  1:15 ` William S. Morgart
  1 sibling, 0 replies; 10+ messages in thread
From: William S. Morgart @ 2008-01-10  1:15 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

I had the same thing happen ... ide dma is turned off and when I turned
it by adding dmaon near the beginning of either termrc or cpurc (or both
if you have both a cpu server and terminal) interrupts returned to a
more normal level and performance is as it should be

On Wed, 2008-01-09 at 18:05 +0100, a.verardo@tecmav.com wrote:
> ----- Original Message -----
> Da : erik quanstrom <quanstro@quanstro.net>
> A : 9fans@cse.psu.edu
> Oggetto : Re: [9fans] Fossil+venti install problem
> Data : Wed, 9 Jan 2008 09:42:01 -0500
> 
> > have you considered installing just fossil.
> > you could then install fossil+venti on a seperate disk/
> > partition on your running system and see if that hangs
> > up.
> > 
> > - erik
> 
> Yes. I could, but a reasonable hypothesis is that fossil
> issue
> a lot of requests caused by a venti unexepected behaviour.
> Probably the install procedure is ok.
> Could it be an endless backup due to a wrong (interpretation
> of the)
> sys clock ?
> 
> adriano


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09 17:46   ` john
@ 2008-01-09 18:01     ` Gabriel Diaz Lopez de la Llave
  0 siblings, 0 replies; 10+ messages in thread
From: Gabriel Diaz Lopez de la Llave @ 2008-01-09 18:01 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

hello

unless the time difference is high there should be no problems iirc  
Russ answered this on a thread about timezone changes and venti, the  
error message was about arena creation time.

slds.


gabi





On Jan 9, 2008, at 18:46, john@csplan9.rit.edu wrote:

>> On Wed Jan  9 12:05:15 EST 2008, a.verardo@tecmav.com wrote:
>>> Yes. I could, but a reasonable hypothesis is that fossil
>>> issue
>>
>> hypothesis is great.  but unless you try something to limit the
>> possbilities, hypothesis is not useful.
>>
>>> a lot of requests caused by a venti unexepected behaviour.
>>> Probably the install procedure is ok.
>>> Could it be an endless backup due to a wrong (interpretation
>>> of the)
>>> sys clock ?
>>
>> why would your clock be different from everone elses?
>>
>> - erik
>
> Doesn't venti get unhappy when your current time is earlier
> than the last-written block or something like that? I thought
> problems could also arise with fossil/venti if you had two
> timesyncs running.
>
> I of course never have any problems because my systems are
> always configured PERFECTLY ;)
>
> John


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09 17:35 ` erik quanstrom
@ 2008-01-09 17:46   ` john
  2008-01-09 18:01     ` Gabriel Diaz Lopez de la Llave
  0 siblings, 1 reply; 10+ messages in thread
From: john @ 2008-01-09 17:46 UTC (permalink / raw)
  To: 9fans

> On Wed Jan  9 12:05:15 EST 2008, a.verardo@tecmav.com wrote:
>> Yes. I could, but a reasonable hypothesis is that fossil
>> issue
> 
> hypothesis is great.  but unless you try something to limit the
> possbilities, hypothesis is not useful.
> 
>> a lot of requests caused by a venti unexepected behaviour.
>> Probably the install procedure is ok.
>> Could it be an endless backup due to a wrong (interpretation
>> of the)
>> sys clock ?
> 
> why would your clock be different from everone elses?
> 
> - erik

Doesn't venti get unhappy when your current time is earlier
than the last-written block or something like that? I thought
problems could also arise with fossil/venti if you had two 
timesyncs running.

I of course never have any problems because my systems are
always configured PERFECTLY ;)

John


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09 17:05 a.verardo
@ 2008-01-09 17:35 ` erik quanstrom
  2008-01-09 17:46   ` john
  2008-01-10  1:15 ` William S. Morgart
  1 sibling, 1 reply; 10+ messages in thread
From: erik quanstrom @ 2008-01-09 17:35 UTC (permalink / raw)
  To: 9fans

On Wed Jan  9 12:05:15 EST 2008, a.verardo@tecmav.com wrote:
> Yes. I could, but a reasonable hypothesis is that fossil
> issue

hypothesis is great.  but unless you try something to limit the
possbilities, hypothesis is not useful.

> a lot of requests caused by a venti unexepected behaviour.
> Probably the install procedure is ok.
> Could it be an endless backup due to a wrong (interpretation
> of the)
> sys clock ?

why would your clock be different from everone elses?

- erik


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
@ 2008-01-09 17:05 a.verardo
  2008-01-09 17:35 ` erik quanstrom
  2008-01-10  1:15 ` William S. Morgart
  0 siblings, 2 replies; 10+ messages in thread
From: a.verardo @ 2008-01-09 17:05 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs


----- Original Message -----
Da : erik quanstrom <quanstro@quanstro.net>
A : 9fans@cse.psu.edu
Oggetto : Re: [9fans] Fossil+venti install problem
Data : Wed, 9 Jan 2008 09:42:01 -0500

> have you considered installing just fossil.
> you could then install fossil+venti on a seperate disk/
> partition on your running system and see if that hangs
> up.
> 
> - erik

Yes. I could, but a reasonable hypothesis is that fossil
issue
a lot of requests caused by a venti unexepected behaviour.
Probably the install procedure is ok.
Could it be an endless backup due to a wrong (interpretation
of the)
sys clock ?

adriano


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
  2008-01-09 14:28 a.verardo
@ 2008-01-09 14:42 ` erik quanstrom
  0 siblings, 0 replies; 10+ messages in thread
From: erik quanstrom @ 2008-01-09 14:42 UTC (permalink / raw)
  To: 9fans

have you considered installing just fossil.
you could then install fossil+venti on a seperate disk/
partition on your running system and see if that hangs
up.

- erik


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [9fans] Fossil+venti install problem
@ 2008-01-09 14:28 a.verardo
  2008-01-09 14:42 ` erik quanstrom
  0 siblings, 1 reply; 10+ messages in thread
From: a.verardo @ 2008-01-09 14:28 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs


----- Original Message -----
Da : erik quanstrom <quanstro@quanstro.net>
A : 9fans@cse.psu.edu
Oggetto : Re: [9fans] Fossil+venti install problem
Data : Wed, 9 Jan 2008 08:11:04 -0500

> > On Dec 27 I wrote about a problem I had installing a
> > fossil+venti config using the ISO image downloaded on
> > Dec 22:
> > 
> > * On several HW setup, from P500 to P3000, the
> > * install goes ok but just after the boot from HD
> > * starts an endless activity (> 20 h ?!) with a steady
> > * 100% cpu load and periodic (1 sec seek, 3 sec. pause)
> > * disk load. Then the machine hangs.
> > 
> > The same behaviour has also been observed in the
> > University lab by some students working on a thesis
> > about Plan9. 
> > Is there other 9fan(s) who had the same problem ?
> > 
> > a.v.
> 
> i hesitate to answer because i have not seen such a
> problem nor do i run venti or fossil.
> 
> i haven't any answers, but perhaps these questions would
> help narrow things down (assuming of course that you
> haven't hit an actual venti bug):
> 
> 1. what kind of drives are you using; what hba?  have you
> enabled dma? 2. which processes are accumulating cpu time?
> 3. what is the interrupt load, memory usage? ("stats
> -lsef")
> 
> - erik
About 1. 2. and 3. I've no precise info now. The CPU is
mainly used by fossil/venti and the load is high enough to
freeze
rio almost totally (the mouse cursor moves every 5-6 sec on
a P3000).
Have been used PCs very different in age/disks/ram etc,
always with the same result.

Usually, just after the first boot, there is a strong disk
activity
only for a few minutes and with a reasonable rio response.
After that
the system get idle it is possible proceed with the
configuration. 

This is exactly the behaviour with all the other ISOs,
included
the one downloaded on Jan-07.

In all tests has been repeated exactly the same "default"
install
procedure without seeing any message error.

adriano


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [9fans] fossil+venti install problem
@ 2007-12-27 11:10 a.verardo
  0 siblings, 0 replies; 10+ messages in thread
From: a.verardo @ 2007-12-27 11:10 UTC (permalink / raw)
  To: 9fans

Hi, all.

I'm trying to install a fossil+venti file server using
the ISO dowloaded on Dec 22.

On several HW setup, from P500 to P3000, the
install goes ok but just after the boot from HD
starts an endless activity (> 20 h ?!) with a steady
100% cpu load and periodic (1 sec seek, 3 sec. pause)
disk load. Then the machine hangs.

Absolutely no problems on the same HW configurations
with the 16-Jan-07 ISO.

Any ideas ?

Thanks in advance.

a.verardo


^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2008-01-10  1:15 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-01-09  8:17 [9fans] Fossil+venti install problem a.verardo
2008-01-09 13:11 ` erik quanstrom
  -- strict thread matches above, loose matches on Subject: below --
2008-01-09 17:05 a.verardo
2008-01-09 17:35 ` erik quanstrom
2008-01-09 17:46   ` john
2008-01-09 18:01     ` Gabriel Diaz Lopez de la Llave
2008-01-10  1:15 ` William S. Morgart
2008-01-09 14:28 a.verardo
2008-01-09 14:42 ` erik quanstrom
2007-12-27 11:10 [9fans] fossil+venti " a.verardo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).