yes maybe, but microsoft can't package all the info and experience needed to make it possible for a msce admin to setup a cluster. It is really much more complicated than that.
It's taken years of development to get the linux cluster packages to where they are now, MS can't just step in and take over with their poorly written software.
So, they plan to "beat linux on value" in the clustering area because they realize that they really can't beat the price.
But -- how would a 100 node microsoft cluster have any better value than the same cluster running some linux clustering sw? The microsoft system would be around 100 times more expensive, and the licensing would be outrageous.
Imagine you want to add 20 nodes to your cluster. With linux -- no problem, cable it up and go. With microsoft, well, you probably have to get some more licenses, and another 20 copies of windows to install. That's around $3500 just for the os software.
And finally, there are lots of linux clustering installations running today, and many of those have been using clusters for years and have a history of upgrades and improvements. I really doubt these people will be interested in switching to a microsoft monolithic cluster.
check out the article at http://www.fastcompany.com/online/01/frog.html
it actually shows the opposite of the frog boiling myth. makes sense, really. if you put a frog in boiling water, it will be severely injured right away and probably won't be able to jump out. Whereas a frog in cold water will get bored and jump out before long.:)
from the article...
How did our expert interpret this triumph of science? "There are certain cases where gradual change is almost preferred," Hofman commented. "The change myth assumes a very narrow view of people. If frogs can do it, people definitely can."
I wonder if the same applies to people and DRM software??
From what I've seen, the places with 1000 machines and 5 admins write their own scripts and monitors.
The places with 100 machines and 30 admins try to buy something like Tivoli, CA/Unicenter, or this new Sun montrosity because their Management won't hire people with experience. I think they think it's cheaper that way.
isn't this a lot like the tv commerical public service announcement where a student asks the librarian for some books, she says 'Oh! These are not allowed anymore.' and someone in dark glasses and a dark coat comes up behind the student and helps escort him away to be interviewed.
The end of the psa says something like 'what if you weren't in America?'
anyone know more details about this (rather poorly described) commercial?
This is not the way Founding Fathers wanted us to vote. Although voting has become marginalized and something that few people do anymore, it still remains a very important and solemn duty. South Florida has a lot of emotional people, and if the inherent inaccuracy of the existing voting machines can help to offset the effect of the incorrectly-cast votes, then I am in favor of it.
you're joking, right? the votes should be miscounted because people don't know any better and they really wanted to vote for someone else but got tricked?
the founding fathers were the type of people who wanted to get the vote to people and out of the hands of the people who knew better and wanted to decide for the rest of us. that's what the whole war was about.
But what would really cheese off the founding fathers is the fact that more than 80% of the eligible voters simply ignore the vote completely.
That leaves the power of the vote to those who do go to the polls multiplied by thousands of times.
Re:Because Linus says dump isn't reliable.
on
Linux Backups Made Easy
·
· Score: 3, Informative
he says right there in the linked article that dump can't reliably back up the filesystem because of the kernel filesystem caching, and that future kernel development is headed further in that direction, so you might as well not depend on dump.
seems pretty reasonable to me, go ahead and use dump if you like though
... Imagine working in an office/cubicle with 32 keyboards and 64 mice, rj45 and rj11 jacks everwhere, throw in some extra pc cases to fill every inch under your desk -- with only one of each that actually works
You must know the guy who set up our office network
A single athalonMP 2200 was smokin my dual xeon setup!
I'd guess that the two processors were spending a lot of time waiting on the single disk. One may be writing out data while the other is trying to read. Each would get a little bit, then the disk goes to the other.
You'd have a lot of overhead while the disk thrashes around. It would be pretty noisy but it would sound like a lot was getting done.
With the single processor, there's more read and written at once and the disk doesn't thrash as quickly to other processes.
If you split the load onto more drives, you may see a huge increase in speed. You could simply mount one drive with the input data, and mount the other for the output directory. That would tend to keep the head movement for each drive localized much more than with just one disk.
with the disk bottleneck gone, you should see double or better speed improvements, unless there really is a drastic problem with intel p4 chips
Madly trying to get a frame to disk? You've never rendered anything in your life... have you.
you know, for a second I thought, 'whoops!'
but then I realized again that the original message said the fast rendering took 14 hours, so it's obvious that they are making a set of frames to view at full speed later on.
this sounds like you have a bottleneck in the disk io area more than a problem from lack of cpu power. it's probably trying madly to get each frame onto disk and get the data for the next off disk.
you could do something simple like software raid with the ide controllers on the motherboard and a couple disks used in parallel. that would be fairly cheap to test out and would probably nearly double your speed.
for more speed, switch to an ide raid controller and you can really get some throughput.
but it's really not as simple as saying that the p4 from intel is poor. there are many things to deal with in system performance tuning, and just getting the biggest/newest processor doesn't make it the fastest system.
And I think it's ironic that some/.'ers think this exploit is such a trivial one to pull off that it makes https:// worthless.
nobody said it was trivial, but the existence of this exploit makes the validity of a signed certificate much less strong.
The whole point of https and verisign was that it guarentees, with the best available digital cryptography, that the server you are talking to is really who it says.
This is the whole purpose of companies like Thawte and Verisign. I'll be that they think it's a serious issue.
Why bother with verisign at all if any cert can be used to sign another?
I still think it's ironic that microsoft has done this to Verisign.
I think it's ironic that MS has pushed netscape and the rest out of the browser market, and has managed to make the purpose of ssl worthless since most of the browers out there will not ever be updated.
Did netscape around 1996/1997 have this bug when it was competing against microsoft/explorer?
What about those other older browsers which have gone by the wayside?
I think you really give microsoft credit for things they aren't even doing. What sort of 'testing' could they be doing if they constantly release such poor code?
And further, what good is a patch if nobody can install it because it's being tested?
"The cows are out of the barn." The idea of ssl ensuring the identity of the server on the remote end is really key to the safety of ecommerce and sending credit card numbers over the net. With this bug, you might as well just throw out https for authentication purposes.
Slashdot Mirror
yes maybe, but microsoft can't package all the
info and experience needed to make it possible for
a msce admin to setup a cluster. It is really
much more complicated than that.
It's taken years of development to get the linux
cluster packages to where they are now, MS can't
just step in and take over with their poorly written
software.
also, people do pay big bucks for linux software
when it is actually worth the money.
Oracle would be a good example of that.
So, they plan to "beat linux on value" in the
clustering area because they realize that they
really can't beat the price.
But -- how would a 100 node microsoft cluster have
any better value than the same cluster running some
linux clustering sw? The microsoft system would
be around 100 times more expensive, and the
licensing would be outrageous.
Imagine you want to add 20 nodes to your cluster.
With linux -- no problem, cable it up and go.
With microsoft, well, you probably have to get
some more licenses, and another 20 copies of
windows to install. That's around $3500 just
for the os software.
And finally, there are lots of linux clustering
installations running today, and many of those
have been using clusters for years and have a
history of upgrades and improvements. I really
doubt these people will be interested in
switching to a microsoft monolithic cluster.
More and more, microsoft is getting desperate.
Apparently, every finite sequence of bits will appear in pi *somewhere*
Really? That sounds like something begging a proof
or a url to a webpage with a proof.
It sounds unlikely that every finite sequence is
in there, but maybe its possible.
"First they ignore you, then they laugh at you, then they fight you, then you win. "
:)
--Ghandi
This quote seems to fit the situation with MS and
linux more and more.
check out the article at http://www.fastcompany.com/online/01/frog.html
:)
it actually shows the opposite of the frog boiling
myth. makes sense, really. if you put a frog
in boiling water, it will be severely injured
right away and probably won't be able to jump out.
Whereas a frog in cold water will get bored and
jump out before long.
from the article...
How did our expert interpret this triumph of science? "There are certain cases where gradual change is almost preferred," Hofman commented. "The change myth assumes a very narrow view of people. If frogs can do it, people definitely can."
I wonder if the same applies to people and DRM
software??
From what I've seen, the places with 1000 machines
and 5 admins write their own scripts and monitors.
The places with 100 machines and 30 admins try to
buy something like Tivoli, CA/Unicenter, or this
new Sun montrosity because their Management won't
hire people with experience. I think they think
it's cheaper that way.
well, it will never work. sun likes to pretend
a lot of things, and this is just one of the
funnier ones in awhile.
I agreee -- but the connection I get is plenty
fast enough all the time on 802.11b. I've
even used it to update my gentoo laptop.
--Rodney
isn't this a lot like the tv commerical public
service announcement where a student asks the
librarian for some books, she says 'Oh! These
are not allowed anymore.' and someone in dark
glasses and a dark coat comes up behind the
student and helps escort him away to be interviewed.
The end of the psa says something like 'what if you
weren't in America?'
anyone know more details about this (rather poorly
described) commercial?
...
unless you carry around a CD everywhere you go, you're probably relying on passwords in the end anyway.
I've got a usb keychain filesystem with my
keys on it for just that purpose.
This is not the way Founding Fathers wanted us to vote. Although voting has become marginalized and something that few people do anymore, it still remains a very important and solemn duty. South Florida has a lot of emotional people, and if the inherent inaccuracy of the existing voting machines can help to offset the effect of the incorrectly-cast votes, then I am in favor of it.
you're joking, right? the votes should be miscounted
because people don't know any better and they
really wanted to vote for someone else but got
tricked?
the founding fathers were the type of people who
wanted to get the vote to people and out of the
hands of the people who knew better and wanted to
decide for the rest of us. that's what the whole
war was about.
But what would really cheese off the founding
fathers is the fact that more than 80% of the
eligible voters simply ignore the vote completely.
That leaves the power of the vote to those who
do go to the polls multiplied by thousands of
times.
he says right there in the linked article that
dump can't reliably back up the filesystem
because of the kernel filesystem caching, and that
future kernel development is headed further in that
direction, so you might as well not depend on dump.
seems pretty reasonable to me, go ahead and use
dump if you like though
Google cache
/.ed . thank you!
yes -- that was a refreshing change from the
usual postings where the page is
I've wondered if Tivo qualifies for the electric
monk from the Dirk Gently Holistic Detective book
I can't count the number of shows that the tivo
decided I should record which I've never really
watched before the space was reused for another
show.
granted, the Tivo doesn't ride a horse
You must know the guy who set up our office network
it doesn't take all that long with the cd drives
today. you probably used something much slower
to do it the first time.
it only takes a few minutes with a good cpu and
fast drive
A single athalonMP 2200 was smokin my dual xeon setup!
I'd guess that the two processors were spending
a lot of time waiting on the single disk. One
may be writing out data while the other is trying
to read. Each would get a little bit, then the
disk goes to the other.
You'd have a lot of overhead while the disk
thrashes around. It would be pretty noisy but
it would sound like a lot was getting done.
With the single processor, there's more read
and written at once and the disk doesn't thrash
as quickly to other processes.
If you split the load onto more drives, you
may see a huge increase in speed. You could
simply mount one drive with the input data,
and mount the other for the output directory.
That would tend to keep the head movement for
each drive localized much more than with just
one disk.
with the disk bottleneck gone, you should see
double or better speed improvements, unless there
really is a drastic problem with intel p4 chips
Madly trying to get a frame to disk? You've never rendered anything in your life... have you.
you know, for a second I thought, 'whoops!'
but then I realized again that the original
message said the fast rendering took 14 hours,
so it's obvious that they are making a set of
frames to view at full speed later on.
so, take your smugness and go away.
At least I sign my posts.
this sounds like you have a bottleneck in the
disk io area more than a problem from lack of cpu
power. it's probably trying madly to get each
frame onto disk and get the data for the next off
disk.
you could do something simple like software raid
with the ide controllers on the motherboard and
a couple disks used in parallel. that would be
fairly cheap to test out and would probably
nearly double your speed.
for more speed, switch to an ide raid controller
and you can really get some throughput.
but it's really not as simple as saying that the p4
from intel is poor. there are many things to
deal with in system performance tuning, and
just getting the biggest/newest processor doesn't
make it the fastest system.
And I think it's ironic that some
nobody said it was trivial, but the existence of
this exploit makes the validity of a signed
certificate much less strong.
The whole point of https and verisign was that it
guarentees, with the best available digital cryptography,
that the server you are talking to is really who
it says.
This is the whole purpose of companies like
Thawte and Verisign. I'll be that they think
it's a serious issue.
Why bother with verisign at all if any cert can
be used to sign another?
I still think it's ironic that microsoft has
done this to Verisign.
I think it's ironic that MS has pushed netscape and
the rest out of the browser market, and has managed
to make the purpose of ssl worthless since most
of the browers out there will not ever be updated.
Did netscape around 1996/1997 have this bug when
it was competing against microsoft/explorer?
What about those other older browsers which have
gone by the wayside?
hmm...
things they aren't even doing. What sort of
'testing' could they be doing if they constantly
release such poor code?
And further, what good is a patch if nobody can
install it because it's being tested?
"The cows are out of the barn." The idea of ssl
ensuring the identity of the server on the remote
end is really key to the safety of ecommerce and
sending credit card numbers over the net. With
this bug, you might as well just throw out https
for authentication purposes.
I have never, ever, thought of fortran and xml at the same time. yikes!
just think about trying to parse those angle
brackets in fortran
hey I get to post my favorite fortran joke
about variable naming.
"God is real; unless declared integer"
my other favorite math joke is
"what's purple and commutes?"