Quantity... But is the quality still there? On the first iteration, millions of lines of code means hundreds of thousand bugs. Does it have any useful features that the one hundred thousand lines of code version of the software didn't have, or does it just have a new nice spash screen and more bugs?
Is it actual new code, or mainly just reused and repackaged code?
(btw: large compared to what, the laptop? the desktop?)
(abou the metaphor: huh? there is no such thing as too large)
"too expensive"
Buy a slightly older refurbished model and you'll be able to afford one too (also valid for the metaphor...)
"too slow"
Huh? Just don't try to play those 3D games on them and you're fine (for the metaphor, just stay away from the fourcylinder kittens).
"and use too much energy."
Palms last up to 5 times longer than PocketWindos PDAs. 10-15 hours of continous use is pretty common. If your flight is longer than that, then tell your boss to send somebody else.
And yes, as you mention, _during_ the meeting, the PDA is the ideal boredom coping tool.
(for the metaphor: many sedans get similar mpg as many SUVs).
Crap? Wrong? Its my experience, that's all. If you read back my posting then you'll see that I'm telling you _my_ experience with ext2 and ntfs, not yours. Perhaps I've been unlucky with ntfs and lucky with ext2, or perhaps it's the other way around for you.
AFAIK, the 'journaling' in NTFS isn't ordered journaling. It keeps a metadata file which get a log of the file actions to allow rollback. But, acording to this page, that file gets written to only once every few seconds. So when you crash, the log may be a few seconds old, so you may end up after reboot with an apparently 'clean' filesystem but some files will be old, missing their latest modifications. ext3 has a mode like that too, 'writeback', but also has the 'journal' and 'ordered' modes that don't have that problem.
The choking of NT for the fab was a result of fragmentation in the filesystem... Another article about the same fab mentioned that (sorry, have no link, but others are reporting it too). That's not what I what I call reliable and robust: Eventually it will fragment too much to perform, resulting in more downtime. It doesn't stay as fast as it was when I built and benchmarked the system... Plus things in the past like this and this don't help at all on the confidence front.
I had many hardware failures on active loaded linux boxes, failing power supplies, mainboards, etc, never lost data as the result (except when a disk completely died without raid, and the disk wouldn't spinup...). The boxes just keep going until the hardware dies. That's what I call reliability and robustness.
About the 'stays fast under any load', I mean cases like the 100% CPU load on fragmented ntfs volumes, as reported here, which cause the system to be unresponsive, resulting in network timeouts... Or things like this, where the given workaround basically means you can't have a lot of small files on your disk...
I've lost more data on a live ntfs system than I've ever lost on a powerfailure with an active ext2... Sure the e2fsck often printed lots of messages, but that was just for the large amount of metadata that remained in the cache when the brute poweroff happened, and was usually fixed without losing data except some recent log entries and other such growing or open files. Added to that the journal of ext3 not virtually guarantees no data loss (and eliminates the extreme fsck wait).
While ntfs needs to be more resilient agains crashes for obvious reasons, even in the crash resiliency respect I've never been mistreated by ext2. It's not a cleanly designed filesystem, it's a stack of features upon a hack. ntfs is just a hack on top of vfat, which is a hack on top of fat. Reminds me of Duck Tape Repairs(tm). It's always in the consistent state of of disarray (In this IBM factory, they luckily found that out ahead of time, with windows machines thrashing and needing defragmentations or reinstalls while Linux just chugged along (I think there was a/. story about this fab, but couldn't find it)). The fact that ntfs doesn't have or use the 'filesystem dirty' state isn't something to be proud of, it results in filesystem errors that remain undetected until the whole thing falls apart. On windows, it's wise to force a full fsck after a crash even if the box seems to be back up.
From a system administration perspective, I want a filesystem that keeps its data and stays fast under any load. About losing a whole server, or a random part of it. Actually, I'd rather lose the whole server and switch to the backup while restoring the tapes than get it started back up and thinking I kept all the data, only to realize that I'm missing part two weeks later just after the backups got overwritten...
You probably have a hardware problem on the laptop (compiling the kernel is a known hardware problem finder because of its unique system load). On the celeron, you may have a missing or wrong supporting package, for exampel the/usr/include/linux directory may not be the correct one.
The RH8 install tree doesn't give you the.config configuration file that you need to build a RH8 compatible kernel (unless you copy it manually, see the linuxexperience.com link below), so the P4 kernel panic could be either a misconfigured kernel (something like missing 'initrd support'), or something misconfigured in/etc/lilo.conf (wrong root partition).
Did you check the NTFS development webpage. It has a special page for RedHat. Also this linkshould be helpful to het ntfs on redhat. It contains a step-by-step explanation.
If you just want ntfs support, you don't need a new kernel. All you need is the ntfs.o module that is compatible with your kernel. Check if you have/lib/modules/*/fs/ntfs.o or/usr/src/linux/fs/ntfs/ntfs.o. If you do, try to insmod it. If you don't, you don't need to do 'make bzImage' or 'make zImage' to ge the module, a 'make modules' should give you the ntfs.o (if you configured it with 'make xconfig' or 'make menuconfig' to make the ntfs support module). If you get compile errors doing 'make modules', for any reason, just swith off compilation of any modules except ntfs (look for 'BLABLA=m' lines in the.config to find out what is configured to be compiled as module).
I use Debian, so I can't verify it for RH, but I've had no problems getting a working ntfs support module for Debian.
"Take the exact goddamn makefile and code tree, and it's compiling on my P4 desktop, but not the Celeron fileserver OR the PII 266 laptop."
Golden tip of the day: The Makefile is not the kernel configuration file, it's the.config file that configures it. Note you need to do a 'make xconfig', or 'make menuconfig', or 'make oldconfig' if you're using a 'clean' unpacked kernel tarball and want to apply the.config file you just copied into it.
What is the error you get? Have you investigated further or did you only try the blind approach of 'lets try this on another machine'?
Sounds like you have a "CONFIG_MPENTIUM4=y" in your.config
Or something similar in your gcc, binutils, or libc6-dev, or you're simply running out of memory/swap.
BTW, none of my windows machines can recompile their kernel.
The point he was making that in windows it's often not documented, so it often ends up with trial and error, hours of kb searches, or just reinstalling with fingers crossed. In Linux, you have the source and the configuration files are human readable, you just need to find the beginning of the yellow brick road and follow it.
"bringing up Runlevels,"
Start at/etc/inittab and follow the rail of scripts, first the script on the 'si:' line, then look at the 'default:' line, and follow the 'l?:' line with '?' the runlevel. You'll probably find most your start and stop scripts in/etc/init.d, and/etc/rcS.d with links from/etc/rc?.d
Most other configurations are in/etc, and per user ones in '.*' (hidden) files or directories of the user's home directory.
Sendmail problems? Try postfix, you'll love it. Easier to configure, easier to understand, and better security track record. btw, configuration is in/etc/postfix... duh.
Got a kernel panic and it's not because youre using the 2.5.x unstable kernels? -> Most probably hardware that is breaking down.
"Further, NTFS is extremely robust and resilient."
In my experience not in the speed sense. Actually, not on any front when compared with ext3. Even when compares with ext2, NTFS is a dog. Where do you get the idea that ext2 is 'horrid'? You must be confusing with last year's reiserfs versions. The only disadvantage of ext2 was the long fsck on large partitions, which was fixed with journaling. The other disadvantage, slowness with many (>5000) files in one subdirectory is worse with NTFS, and will be fixed (may already be fixed) in one of the upcoming kernel releases anyway.
My experience is, more reliable (less crashes and lost work), and much less annoyances dealing with the IT department to get licenses for the desktop, laptop, etc. The only drawbacks I've encountered is first a much slower startup time, but that is only for the first document, and second it's very close but still not 100% capable of reading and writing MS formats with no formatting mistakes.
Just this week, I had to rush out a large document for a customer. But it needed some final preparations before I could send it off. I opened it, in word2000 (the company standard...), started editing, moving things around, etc, and poof. An error box pops up and bye bye word2000. Restarting word, document not recovered, it just showed a new empty document. Oh great! And this is on win2000, so don't blame it on winme. Anyways, here is the switch ad: So I ditched word2000 and used OpenOffice and finished the document in time without any problems.
Where do all these myths come from that MSOffice is more productive than the other office suites? If it's not MS FUD, then it's got to be inexperience using any of the other office suites out there.
If it's the only ice cream you've ever had, you'll think it's the greatest... even if it's not.
With this news, a wireless display for windows XP only, and the new MS tables PCs, now I suddenly realize what steve meant a couple of weeks ago with outsmarting Linux.
Of course... It will work only until the porting projects get underway...
It's a cluster, so I can imagine the nodes can boot individually, in parallel. Plus I can imagine the system never goes down as a whole, just some nodes may go down when parts break or other maintenance. 1 bootup per lifetime...
Perhaps the boot speed is limited by the ramp-up speed of the local power plant.
"The tricky part would be the clock pulses. In order to keep the data integrity, you need to have both bachines on the same clock. The easy way would be to take the crystal from one motherboard and wire it to the other. Same crystal, same clock pulse."
With the frequencies and distances involved, I doubt that that will work. I think the tranmission line delay of the wires will mess it up. I suspect they have some sort of dual-port buffer in the cards to allow multiple clock domains.
What I'm wondering is: Where can I get those cards and how much do they cost?
They are taking power from the tides? The tides are generated from the gravitational pull of the moon. Taking power from it reduces the orbit of the moon, inevitably making it crash into the earth. Doom Doom Panic Panic.
I wonder how many exawatt-years that would be until it gets that far though...
Slashdot Mirror
"millions of lines of code"
Quantity... But is the quality still there? On the first iteration, millions of lines of code means hundreds of thousand bugs. Does it have any useful features that the one hundred thousand lines of code version of the software didn't have, or does it just have a new nice spash screen and more bugs?
Is it actual new code, or mainly just reused and repackaged code?
What are you talking about? The site is up and running. These guys obviously are /. resistant.
I switched to buying mail-order off the Internet because of those annoying questions.
Doh!
(hint: mailorder, address... bingo)
"too large",
Try a Clie.
(btw: large compared to what, the laptop? the desktop?)
(abou the metaphor: huh? there is no such thing as too large)
"too expensive"
Buy a slightly older refurbished model and you'll be able to afford one too (also valid for the metaphor...)
"too slow"
Huh? Just don't try to play those 3D games on them and you're fine (for the metaphor, just stay away from the fourcylinder kittens).
"and use too much energy."
Palms last up to 5 times longer than PocketWindos PDAs. 10-15 hours of continous use is pretty common. If your flight is longer than that, then tell your boss to send somebody else.
And yes, as you mention, _during_ the meeting, the PDA is the ideal boredom coping tool.
(for the metaphor: many sedans get similar mpg as many SUVs).
Thanks for the tip. I've been looking for a good free palm app to do that!
Crap? Wrong? Its my experience, that's all. If you read back my posting then you'll see that I'm telling you _my_ experience with ext2 and ntfs, not yours. Perhaps I've been unlucky with ntfs and lucky with ext2, or perhaps it's the other way around for you.
AFAIK, the 'journaling' in NTFS isn't ordered journaling. It keeps a metadata file which get a log of the file actions to allow rollback. But, acording to this page, that file gets written to only once every few seconds. So when you crash, the log may be a few seconds old, so you may end up after reboot with an apparently 'clean' filesystem but some files will be old, missing their latest modifications. ext3 has a mode like that too, 'writeback', but also has the 'journal' and 'ordered' modes that don't have that problem.
The choking of NT for the fab was a result of fragmentation in the filesystem... Another article about the same fab mentioned that (sorry, have no link, but others are reporting it too). That's not what I what I call reliable and robust: Eventually it will fragment too much to perform, resulting in more downtime. It doesn't stay as fast as it was when I built and benchmarked the system... Plus things in the past like this and this don't help at all on the confidence front.
I had many hardware failures on active loaded linux boxes, failing power supplies, mainboards, etc, never lost data as the result (except when a disk completely died without raid, and the disk wouldn't spinup...). The boxes just keep going until the hardware dies. That's what I call reliability and robustness.
About the 'stays fast under any load', I mean cases like the 100% CPU load on fragmented ntfs volumes, as reported here, which cause the system to be unresponsive, resulting in network timeouts... Or things like this, where the given workaround basically means you can't have a lot of small files on your disk...
Not only that. It's lauching at me!
Or before you post, learn geography. The map is of the region of Twente, the University is just north of Enschede in the East.
I've lost more data on a live ntfs system than I've ever lost on a powerfailure with an active ext2... Sure the e2fsck often printed lots of messages, but that was just for the large amount of metadata that remained in the cache when the brute poweroff happened, and was usually fixed without losing data except some recent log entries and other such growing or open files. Added to that the journal of ext3 not virtually guarantees no data loss (and eliminates the extreme fsck wait).
/. story about this fab, but couldn't find it)). The fact that ntfs doesn't have or use the 'filesystem dirty' state isn't something to be proud of, it results in filesystem errors that remain undetected until the whole thing falls apart. On windows, it's wise to force a full fsck after a crash even if the box seems to be back up.
While ntfs needs to be more resilient agains crashes for obvious reasons, even in the crash resiliency respect I've never been mistreated by ext2. It's not a cleanly designed filesystem, it's a stack of features upon a hack. ntfs is just a hack on top of vfat, which is a hack on top of fat. Reminds me of Duck Tape Repairs(tm). It's always in the consistent state of of disarray (In this IBM factory, they luckily found that out ahead of time, with windows machines thrashing and needing defragmentations or reinstalls while Linux just chugged along (I think there was a
From a system administration perspective, I want a filesystem that keeps its data and stays fast under any load. About losing a whole server, or a random part of it. Actually, I'd rather lose the whole server and switch to the backup while restoring the tapes than get it started back up and thinking I kept all the data, only to realize that I'm missing part two weeks later just after the backups got overwritten...
Often, banks have safe storage for you to rent. Some banks even give you a slot for free with a premium checking account.
Otherwise, maybe keep a set of (encrypted!) disks in your car, so at least it's out of the house. Time to get a DVD-R drive!
You probably have a hardware problem on the laptop (compiling the kernel is a known hardware problem finder because of its unique system load). On the celeron, you may have a missing or wrong supporting package, for exampel the /usr/include/linux directory may not be the correct one.
.config configuration file that you need to build a RH8 compatible kernel (unless you copy it manually, see the linuxexperience.com link below), so the P4 kernel panic could be either a misconfigured kernel (something like missing 'initrd support'), or something misconfigured in /etc/lilo.conf (wrong root partition).
/lib/modules/*/fs/ntfs.o or /usr/src/linux/fs/ntfs/ntfs.o. If you do, try to insmod it. If you don't, you don't need to do 'make bzImage' or 'make zImage' to ge the module, a 'make modules' should give you the ntfs.o (if you configured it with 'make xconfig' or 'make menuconfig' to make the ntfs support module). If you get compile errors doing 'make modules', for any reason, just swith off compilation of any modules except ntfs (look for 'BLABLA=m' lines in the .config to find out what is configured to be compiled as module).
The RH8 install tree doesn't give you the
Did you check the NTFS development webpage. It has a special page for RedHat. Also this linkshould be helpful to het ntfs on redhat. It contains a step-by-step explanation.
If you just want ntfs support, you don't need a new kernel. All you need is the ntfs.o module that is compatible with your kernel. Check if you have
I use Debian, so I can't verify it for RH, but I've had no problems getting a working ntfs support module for Debian.
Network traffic rrdtool graph.
"Take the exact goddamn makefile and code tree, and it's compiling on my P4 desktop, but not the Celeron fileserver OR the PII 266 laptop."
.config file that configures it. Note you need to do a 'make xconfig', or 'make menuconfig', or 'make oldconfig' if you're using a 'clean' unpacked kernel tarball and want to apply the .config file you just copied into it.
.config
Golden tip of the day: The Makefile is not the kernel configuration file, it's the
What is the error you get? Have you investigated further or did you only try the blind approach of 'lets try this on another machine'?
Sounds like you have a "CONFIG_MPENTIUM4=y" in your
Or something similar in your gcc, binutils, or libc6-dev, or you're simply running out of memory/swap.
BTW, none of my windows machines can recompile their kernel.
The point he was making that in windows it's often not documented, so it often ends up with trial and error, hours of kb searches, or just reinstalling with fingers crossed. In Linux, you have the source and the configuration files are human readable, you just need to find the beginning of the yellow brick road and follow it.
/etc/inittab and follow the rail of scripts, first the script on the 'si:' line, then look at the 'default:' line, and follow the 'l?:' line with '?' the runlevel. You'll probably find most your start and stop scripts in /etc/init.d, and /etc/rcS.d with links from /etc/rc?.d
/etc, and per user ones in '.*' (hidden) files or directories of the user's home directory.
/etc/postfix... duh.
"bringing up Runlevels,"
Start at
Most other configurations are in
Sendmail problems? Try postfix, you'll love it. Easier to configure, easier to understand, and better security track record. btw, configuration is in
Got a kernel panic and it's not because youre using the 2.5.x unstable kernels? -> Most probably hardware that is breaking down.
"Further, NTFS is extremely robust and resilient."
In my experience not in the speed sense. Actually, not on any front when compared with ext3. Even when compares with ext2, NTFS is a dog. Where do you get the idea that ext2 is 'horrid'? You must be confusing with last year's reiserfs versions. The only disadvantage of ext2 was the long fsck on large partitions, which was fixed with journaling. The other disadvantage, slowness with many (>5000) files in one subdirectory is worse with NTFS, and will be fixed (may already be fixed) in one of the upcoming kernel releases anyway.
"I would like to take a look at OOo"
Why not try it out yourself?
My experience is, more reliable (less crashes and lost work), and much less annoyances dealing with the IT department to get licenses for the desktop, laptop, etc. The only drawbacks I've encountered is first a much slower startup time, but that is only for the first document, and second it's very close but still not 100% capable of reading and writing MS formats with no formatting mistakes.
MSOffice _more_ productive? On what planet?
Just this week, I had to rush out a large document for a customer. But it needed some final preparations before I could send it off. I opened it, in word2000 (the company standard...), started editing, moving things around, etc, and poof. An error box pops up and bye bye word2000. Restarting word, document not recovered, it just showed a new empty document. Oh great! And this is on win2000, so don't blame it on winme. Anyways, here is the switch ad: So I ditched word2000 and used OpenOffice and finished the document in time without any problems.
Where do all these myths come from that MSOffice is more productive than the other office suites? If it's not MS FUD, then it's got to be inexperience using any of the other office suites out there.
If it's the only ice cream you've ever had, you'll think it's the greatest... even if it's not.
Still I prefer that over software that is released before the major gaping security holes have been taken out. Security hole of the week... License change of the month. Cost increase of the year. New enlarged start button wooptydoo.
With this news, a wireless display for windows XP only, and the new MS tables PCs, now I suddenly realize what steve meant a couple of weeks ago with outsmarting Linux.
Of course... It will work only until the porting projects get underway...
Since nobody is answering your question: The Top500 supercomputers are ranked by the results of the LinPack benchmark.
It's a cluster, so I can imagine the nodes can boot individually, in parallel. Plus I can imagine the system never goes down as a whole, just some nodes may go down when parts break or other maintenance. 1 bootup per lifetime...
Perhaps the boot speed is limited by the ramp-up speed of the local power plant.
Why does everybody type 'line of site' when it really is 'line of sight'?
Websight.
Somebody just posted a 10mile working link with a juice can and a recycled dish.
"The tricky part would be the clock pulses. In order to keep the data integrity, you need to have both bachines on the same clock. The easy way would be to take the crystal from one motherboard and wire it to the other. Same crystal, same clock pulse."
With the frequencies and distances involved, I doubt that that will work. I think the tranmission line delay of the wires will mess it up. I suspect they have some sort of dual-port buffer in the cards to allow multiple clock domains.
What I'm wondering is: Where can I get those cards and how much do they cost?
Oh no!
They are taking power from the tides? The tides are generated from the gravitational pull of the moon. Taking power from it reduces the orbit of the moon, inevitably making it crash into the earth. Doom Doom Panic Panic.
I wonder how many exawatt-years that would be until it gets that far though...