Sebastian Rother jpberlin.de> writes:
> vnconfig -cK 52527 -S saltfile /dev/sd0k /dev/svnd1c
I think that does synchronous writes, even if you mount
the svnd device async or softdep, which is why it is so
slow.
After losing a hard disc image _file_ to fsck on the filesy-
stem containing the ima
On Mon, Apr 27, 2009 at 12:26:13AM +0200, Sebastian Rother wrote:
> > If the way you do something take too long.
> > Seems like that is a bug.
> > Most likely in the way you are doing it.
> > A lot of things, you can do them wrong and get away with it for a while.
> > Getting away with doing someth
PS: I made a mistake today in the morning (I wrote the e-mail in the
small hours and I was feeling like dead).
I wrote:
"I have an old usb drive encrypted with /dev/arandom (much slower than
zero) which I mount without softdep
I just finished a full backup of 43.98G
It took 1h33min"
That's wro
* Henning Brauer [2009-04-27 10:33]:
> and now for the pf bug. as usual, you did nothing. you accidently
> found some way to crash a box in a specific setup. you did no work at
> all looking where the bug could be or what could trigger it. nothing.
> I ran the command you claimed crashes pf. my bo
* sebastian.rot...@jpberlin.de [2009-04-25
23:48]:
> I wrote Marco personaly, provided all informations and asked if he needs
> further benchmarks or what-so-ever.
did you find the commit between 4.1 and 4.2 or whatever your claim was
where it got slower?
I am sure what the answer is. You did no
* sebastian.rot...@jpberlin.de [2009-04-26
15:28]:
> You could test the svnd on your own BTW because I doubt it's HW related...
.oO how comes you don't receive replies from developers? I have no idea.
--
Henning Brauer, h...@bsws.de, henn...@openbsd.org
BS Web Services, http://bsws.de
Full-Ser
Hello,
I have an old usb drive encrypted with /dev/arandom (much slower than
zero) which I mount without softdep
I just finished a full backup of 43.98G
It took 1h33min
He is doing something wrong.
And you know... if in my work, somebody comes to me and asks for a
favour, I make it, no problem
Sebastian Rother wrote:
>
> ...but I somehow think I know how to use vnconfig.
and it takes too long.
way too long.
Methinks there's something wrong with that logic.
Does the excess time have something to do with bugs in pf?
If so what?
If not, where is the relevance?
Seems like you are being ta
> If the way you do something take too long.
> Seems like that is a bug.
> Most likely in the way you are doing it.
> A lot of things, you can do them wrong and get away with it for a while.
> Getting away with doing something wrong is far from proof that you were
> doing it right.
That's for sure
as annoying on this list as you are.
> -Original Message-
> From: Sebastian Rother [mailto:sebastian.rot...@jpberlin.de]
> Sent: Sunday, April 26, 2009 4:16 PM
> To: t...@servacorp.com
> Subject: Re: svnd is incredible slow... somebody else notice that?
>
> On Sun,
Sebastian Rother wrote:
> A 16GB backup of /home takes more then 10 hrs to restore.
> It's like ataching the device, rsync -av SOURCE:/FOO . and
> wait for 10+
> hours.
That sounds like you are doing something wrong.
And then you come whining here because you do not know how to write to a
disk?
Th
On Sun, Apr 26, 2009 at 1:52 PM, Sebastian Rother
wrote:
> On Sun, 26 Apr 2009 11:37:24 -0500
> Marco Peereboom wrote:
>
>> You are retarded and unable to figure out what is going on. Spouting
>> horeshit as usual. Seriously just go away.
>
> From one retard to another: Go and fix the retarded
Sebastian Rother wrote:
On Sun, 26 Apr 2009 11:37:24 -0500
Marco Peereboom wrote:
You are retarded and unable to figure out what is going on. Spouting
horeshit as usual. Seriously just go away.
From one retard to another: Go and fix the retarded pf code or whatever
except of talkin
sebastian.rot...@jpberlin.de wrote:
> > Bonnie is retarded and proves nothing one way or another.
> Typical KY
> > for masturbation.
>
> Well then simply tell me how to test/benchmark it?
> You could test the svnd on your own BTW because I doubt it's
> HW related...
>
> I asked you serval time
You know what, I actually have figured out a way to make softraid crypto
go faster but I'll hang on to that until I figure out how to license
code in a way that excludes you using it. I might call it the
sebatianrotherisfuckingretardedGPL license. I swear you make my
appetite to hack diminish pre
On Sun, 26 Apr 2009 11:37:24 -0500
Marco Peereboom wrote:
> You are retarded and unable to figure out what is going on. Spouting
> horeshit as usual. Seriously just go away.
>From one retard to another: Go and fix the retarded pf code or whatever
except of talking in such a way to somebody els
On Sun, Apr 26, 2009 at 12:00 PM, Sebastian Rother
wrote:
>> Why exactly does showing benchmark output make you think svnd gets any
>> faster?
>>
>> Do you believe the developers are going to look at your numbers and fix
>> it for you because your numbers show that the cpu sits around doing
>> not
On Sun, Apr 26, 2009 at 06:00:04PM +0200, Sebastian Rother wrote:
> > Why exactly does showing benchmark output make you think svnd gets any
> > faster?
> >
> > Do you believe the developers are going to look at your numbers and fix
> > it for you because your numbers show that the cpu sits around
You are retarded and unable to figure out what is going on. Spouting
horeshit as usual. Seriously just go away.
On Sun, Apr 26, 2009 at 06:00:04PM +0200, Sebastian Rother wrote:
> > Why exactly does showing benchmark output make you think svnd gets any
> > faster?
> >
> > Do you believe the dev
> Why exactly does showing benchmark output make you think svnd gets any
> faster?
>
> Do you believe the developers are going to look at your numbers and fix
> it for you because your numbers show that the cpu sits around doing
> nothing all day?
svnds are used for make release as well
But I'm s
On Sun, Apr 26, 2009 at 03:15:02PM +0200, sebastian.rot...@jpberlin.de wrote:
> > Bonnie is retarded and proves nothing one way or another. Typical KY
> > for masturbation.
>
> Well then simply tell me how to test/benchmark it?
> You could test the svnd on your own BTW because I doubt it's HW rel
Bonnie is retarded and proves nothing one way or another. Typical KY
for masturbation.
On Sun, Apr 26, 2009 at 12:04:04PM +0200, sebastian.rot...@jpberlin.de wrote:
> Well bonnie++ finished...
>
> Setup + bonie++ result + dmesg
>
> The result is HW indipendent for me (compareable results on am
> Bonnie is retarded and proves nothing one way or another. Typical KY
> for masturbation.
Well then simply tell me how to test/benchmark it?
You could test the svnd on your own BTW because I doubt it's HW related...
I asked you serval times to provide me some hints of what you may need
related
Well bonnie++ finished...
Setup + bonie++ result + dmesg
The result is HW indipendent for me (compareable results on am64 too on
another box). I changed the HDD as well (even I doubt it will help) with
no improvement to the writing speed.
Setup:
OpenBSD-Current, i386 (somebody knows why 1GB ra
sebastian.rot...@jpberlin.de wrote:
On Fri, Apr 24, 2009 at 6:12 PM, wrote:
If you've serval houndret GBs that gonna take a lng time.
Also you can not restore a backup quickly because of the uberproor write
performance (it feels like being slower then PIO 3..).
crypto is slow. what else
> Marco Peereboom wrote:
>> You are right about how awful all this stuff is. Man it seems like you
>> should use an os that suits your goals a little better. I have heard
>> that Linux offers awesome performance.
>>
>
>
> based on the manner in which you routinely complain and provide zero
> delive
> On Fri, Apr 24, 2009 at 6:12 PM, wrote:
>> If you've serval houndret GBs that gonna take a lng time.
>> Also you can not restore a backup quickly because of the uberproor write
>> performance (it feels like being slower then PIO 3..).
>
> crypto is slow. what else is new?
I do not talk ab
Marco Peereboom wrote:
You are right about how awful all this stuff is. Man it seems like you
should use an os that suits your goals a little better. I have heard
that Linux offers awesome performance.
based on the manner in which you routinely complain and provide zero
deliverables, i mus
You are right about how awful all this stuff is. Man it seems like you
should use an os that suits your goals a little better. I have heard
that Linux offers awesome performance.
On Apr 24, 2009, at 17:12, sebastian.rot...@jpberlin.de wrote:
I notice it for a while now that SVND is incredibl
On Fri, Apr 24, 2009 at 6:12 PM, wrote:
> If you've serval houndret GBs that gonna take a lng time.
> Also you can not restore a backup quickly because of the uberproor write
> performance (it feels like being slower then PIO 3..).
crypto is slow. what else is new?
> So what other choices
sebastian.rother () jpberlin ! de wrote
> I notice it for a while now that SVND is incredible slow related to WRITE
> SPEED. Also I do see a lot of "biowait" with top related to newfs for
> example.
>
>
> vnconfig -cK -S saltfile /dev/sd0d /dev/svnd1c
> disklabel -E svnd1
> -> a a
> -> r
> -
I notice it for a while now that SVND is incredible slow related to WRITE
SPEED. Also I do see a lot of "biowait" with top related to newfs for
example.
vnconfig -cK -S saltfile /dev/sd0d /dev/svnd1c
disklabel -E svnd1
-> a a
-> r
-> w
-> q
newfs /dev/rsvnd1a
If you've serval houndret GBs t
32 matches
Mail list logo