[ENBD] (no subject)

[HUDZIA Bertrand]

On Wed, 10 Apr 2002 20:16:18 +0200 (MET DST)
"Peter T. Breuer" <ptb@it.uc3m.es> wrote:

> Interesting. It should not be trying to make anything called "config".
> It should do the ./configure, which makes the Makefile in /tmp from
> Makefile.in, and then calls "make config" against that makefile, and
> there's an explict config target there.
> 
> Perhaps you have old remnants of some other compilation there? try
> "make clean" or "rm /tmp/heho/config* /tmp/heho/Make*".

Nope, it comes from the install on the serv , i'll make a new one.

> 
> > But on the client, it manage to compil well. Binary don't complain
> > running on the other.
> 
> These things are not random!

I'll try to make twin install on two other pc in order to have a "clean"
environement test ;) I'll tell u the result as soon as i'll finished ;)

> 
> > > > Both pc run 2.4.18 kernel under a RedHat-7.2 on a pIII 950Mhz.
> > > > serv got 256 Mo RAM and cli have 128 Mo.
> > > 
> > > So these are uniprocessors? On a uniprocessor kernel?
> > 
> > Yes, uniprocessor kernel .
> 
> OK, well, there are no known problems (at the moment!).
> 
> > > It sounds all OK. Would you mind running nbd-test on the device,
> > > with some size limit like "-s 4G""?  The only thing untested about
> > > your setup is the very large size it has. I run pretty much the
> > > same setup but at 4GB.
> > 
> > Test is running , i'll tell u the result tomorow (yes, i'm in France
> > ;).

All the tests ran without any error ...

I'll try tonight to run it with fullsize.

> > 
> > > The more info you can get the better.  What are your
> > > /proc/sys/vm/bdflush settings? Memory overcommit?
> > 

Arf , i'm not very friendly at this time with those "word", what does it
mean ? Why memory "overcommit" ???

Here are my  /proc/sys/vm/bdflush while running the nbd-client :
25      0       0       0       500     3000    80      0       0


> It's probably a bug ;-). I keep forgetting to bump the version number
> in the code. Yes. Well done. Bug corrected ...

;)


> The message says that some minor setup code was run before anybody
> signed the device. This is not surprising, and is harmless.

Ok ;)

I reproduced the crash, but i ran nbd-client with -s , but it didn't
change anything. I manage to get some of the *last* /proc/nbdinfo (
snapshot every 1s) :

	Device a:       Open 
	[a] State:      verify, rw, enabled, plug, last error 0, lives 2, bp 0
	[a] Queued:     +0R/0W curr (check 0R/33W) +64R/64W max
	[a] Buffersize: 262144  (sectors=512, blocks=256)
	[a] Blocksize:  1024    (log=10)
	[a] Size:       36869175KB
	[a] Blocks:     36869175
	[a] Sockets:    1       (*)
	[a] Requested:  24.647M (24.6M) 16.12MR/8.521MW max 4
	[a] Despatched: 24.647M (24.6M) 16.12MR/8.520MW md5 8.52MW (8.52M eq, 0
ne, 0 dn)	[a] Errored:    0       (0)     0+0
	[a] Pending:    4       (4)     0R/4W+0R/132W
	[a] B/s now:    540K    (11.0KR+527KW)
	[a] B/s ave:    3.98G   (13.1MR+6.96MW)
	[a] B/s max:    4.99M   (3.19GR+1.61GW)
	[a] Spectrum:
	[a] Kthreads:   0       (0 waiting/0 running/1 max)
	[a] Cthreads:   0       (-)
	[a] Cpids:      0       (8272)
	Device b-p:     Closed

	Device a:       Open 
	[a] State:      verify, rw, enabled, plug, last error 0, lives 2, bp 0
	[a] Queued:     +1R/1W curr (check 1R/51W) +64R/64W max
	[a] Buffersize: 262144  (sectors=512, blocks=256)
	[a] Blocksize:  1024    (log=10)
	[a] Size:       36869175KB
	[a] Blocks:     36869175
	[a] Sockets:    1       (*)
	[a] Requested:  24.648M (24.6M) 16.12MR/8.522MW max 4
	[a] Despatched: 24.648M (24.6M) 16.12MR/8.521MW md5 8.52MW (8.52M eq, 0
ne, 0 dn)	[a] Errored:    0       (0)     0+0
	[a] Pending:    0       (0)     0R/0W+4R/204W
	[a] B/s now:    817K    (8.00KR+807KW)
	[a] B/s ave:    3.98G   (13.1MR+6.95MW)
	[a] B/s max:    4.99M   (3.19GR+1.61GW)
	[a] Spectrum:
	[a] Kthreads:   0       (0 waiting/0 running/1 max)
	[a] Cthreads:   0       (-)
	[a] Cpids:      0       (8272)
	Device b-p:     Closed

	Device a:       Open 
	[a] State:      verify, rw, enabled, plug, last error 0, lives 2, bp 0
	[a] Queued:     +11R/11W curr (check 11R/47W) +64R/64W max
	[a] Buffersize: 262144  (sectors=512, blocks=256)
	[a] Blocksize:  1024    (log=10)
	[a] Size:       36869175KB
	[a] Blocks:     36869175
	[a] Sockets:    1       (*)
	[a] Requested:  24.648M (24.6M) 16.12MR/8.522MW max 4
	[a] Despatched: 24.648M (24.6M) 16.12MR/8.522MW md5 8.52MW (8.52M eq, 0
ne, 0 dn)	[a] Errored:    0       (0)     0+0
	[a] Pending:    4       (4)     0R/4W+44R/188W
	[a] B/s now:    618K    (27.0KR+589KW)
	[a] B/s ave:    3.98G   (13.1MR+6.94MW)
	[a] B/s max:    4.99M   (3.19GR+1.61GW)
	[a] Spectrum:
	[a] Kthreads:   0       (0 waiting/0 running/1 max)
	[a] Cthreads:   0       (-)
	[a] Cpids:      0       (8272)
	Device b-p:     Closed

Waooo, i've got an average that is more than the max :! 
(but it was there before i started the crash test, doesn't have any link
here :)

The traffic was not very hudge, it was jumping for 0 to 800K. I don't
have the Spectrum value here , duno why .

Note : I ran the crash test on the client and the serv on their own hdd,
it doesn't crashed.

I have one more question , Am I the only one that manage to crash nbd
with that test ? (maybe i am only dealing with bad hardware ????? but
i'll try on other pc today :)

		Bd.