Electricity was a luxury for a long time. Back then, people had been without it all their lives and thus used nothing that depended on it -- wood stoves, oil lanterns, horses and oxen, ice boxes, etc. Fast forward to today and everything runs on electricity. I'd bet a fair number of homes in the US don't have the materials to even start a fire without electricity. Shit, even lighters employ electricity -- piezo element striker (there's nothing to ever wear out like a flint.)
Just like everything else, once everyone has a thing for long enough, they don't know how to do without it. I grewup (well half way +/-) in a world without the computer -- in the 70's computers where the size of builds. I've worked with computers for ~20 years now; I wouldn't know how to function without them today.
HOWEVER, "broadband" is still a luxury and will remain so for many years to come. 56k modems (53k/48k/whatever) are perfect usable, but time consumingly slow. 128K ISDN/144K IDSL is quite sufficient for what 90% of the world does (legally) -- email, web browsing, IM, any number of games, etc. I used 28.8 dialup for ~5 years, and ISDN for ~7 years. I switched to an Earthlink cable modem about a month ago because it's 1/5th the cost -- yes, cost not bps was the reason and earthlink because they provide dialup for free when I'm out of town. (the increase in speed is a nice bonus, tho':-))
[1981 first introduction to a computer -- Tandy TRS Model I, 1984 first computer of my own -- Tandy Color Computer 2.]
"can run" as in VMS will know what to do with the specific hardware. "Alpha", like "PC", does not refer to a single hardware design. There were systems designed to run VMS/Unix and systems designed to run NT. (which mostly boils down to SRM vs. AlphaBIOS.)
"can run" as in VMS will recognize the hardware and go about it's business. That's like installing a BeOS developer release on a Xeon. (that probablly would work. it just wouldn't see anything other than the processor.)
Tru64 panics attmepting to determine the hardware... when it reads the hwdb, it doesn't understand what the cpu board id is and gives up. It's labeled "DIGITAL Server 3000", but DEC renamed the hardware three times before Tru64 existed... According to linux, it's a Noritake/Primo -- which is the correct MB/cpu codename.
PS: I don't care about VMS. I thought it was odd for them to send me VMS when I requested a "hardware update" CD (which is simply an AlphaBIOS and SRM update CD.)
VIM - Vi IMproved 4.2 (1996 June 17, compiled Dec 17 1996 14:55:08)
[cramer:ttya1]dominion:~/[9:16pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 0 Nov 11 21:16 test [cramer:ttya1]dominion:~/[9:16pm]:echo ":wq" | vim test >/dev/null Vim: Warning: Output is not to a terminal Vim: Warning: Input is not from a terminal [cramer:ttya1]dominion:~/[9:16pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 0 Nov 11 21:16 test [cramer:ttya1]dominion:~/[9:17pm]:echo ":wq" | elvis test >/dev/null [cramer:ttya1]dominion:~/[9:17pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 1 Nov 11 21:17 test [cramer:ttya1]dominion:~/[9:17pm]:which elvis /home/cramer/bin/elvis [cramer:ttya1]dominion:~/[ 9:17pm]:elvis --version ELVIS 1.8pl3, by Steve Kirkendall (22 April 1994) ...
Re:How does he handle renaming and erasing in CVS?
on
Home Directory In CVS
·
· Score: 1
RCS magic behind the scenes. (login to the CVS server and mv the RCS file.)
Heh, I have a copy of VMS (or is it MVS?) DEC sent it to me with a bunch of updates for my alpha. Of course, I have no license keys for any of:-) (And I don't think that alpha can run VMS.)
For the free license, the repo itself would have to be public. Open Logging does not send the changes, just a summary of the files changed and comments. And as of a few weeks ago, you couldn't browse the openlogging tree anyway:-) So, openlogging isn't the problem.
And no one has pointed out the obvious: what do you do when the CVS server dies? CVS is not distributed. There is exactly ONE copy of the repo. You cannot restore the repo from a checked out copy as it only carries one version of the file(s). BK is distributed; one works from a clone of the repo that can be used to restore a lost server.
[Personally, I prefer clearcase for my SCM needs. But it's not distributed either. And there's no flavor of it that's even remotely free. Hmm, a home dir as a clearcase dynamic view??? That'd be tricky.]
Well, are they being "rampantly" copied today? HDTV broadcast stations have been around for many years now. (WRAL-HD was the first one to be licensed back in 1998. I've not heard Mr. Goodman rant for even one second about people recording what he's been broadcasting for the past *5* years.)
Answer: No.
Or more accurately, no more so than analog broadcasts.
NetApp has "certified firmware" not "their own"... There's a difference between saying "we like fw v 4.1.0.5" and having your own custom "only given to us, made specifically for us" firmware. Sun (and Apple of the old days) have custom firmware. In fact, the drive ID says it's a sun (or apple) drive. NetApp drives are still seagate/quantum/whomever drives. And they use firmware you could find in retail drives.
NetApp's quality control requires all the drives have the same proven stable firmware so there are no gotcha's down the line. (I have several dozen drives from netapp arrays, btw.) Which means none of their drives run the latest junk from the factory.
[Please note, it is very rare to update the firmware on a hard drive. (or it used to be.) And it's equally difficult to get the update out of the manufacturer.]
I'll look around for the pictures of it before it was in a case. And after. It was sold to a company that handles processing for insurance forms -- it's their fileserver. (We rolled it out to a Chrysler LeBaron, loaded it in the back seat, and they drove it back to NY. [I'm in NC.])
Because they are installing windows (xp home?) on it. I'd also recommend avoiding NTFS on removable media outside of a domain -- bad things happen when one NTFS volume ends up on a different "alien" system.
Rule #1: Servers don't have video hardware - period.
I never could get an answer as to why all the Sun Enterprise servers at Interpath had Creator 3d video hardware in them? Hello, that's a 2000$ waste. (and why do people install all of X11 on systems that will never X applications?) They were web and email servers.
Maybe I'm a purest.:-) Granted, I spent several years on a VT420 attached to a sparc5 clone and a 486dx50 w/o a video card (yeah, it beeped but still continued to boot.)
Well, mirroring (RAID1) is pretty painless on any system. Arrays are simply clones. It's very easy to deal with situations where a clone has dropped out.
For RAID0, a failure means lost data. "Please drive through."
For RAID>1, software based solutions can be a serious pain in the ass. Ok, what's the commands to quiet the bus for drive removal? what's the commands for restoring a drive? What happens when I update the kernel/raid software/etc.? I've had major trouble with linux basically having to bring the system back to the workbench to get RAID reactivated. I've had minor problems with Solaris -- patches breaking things until I read the solution to a non-public bug report. (A public patch breaks things but only contract customers are told how to fix it.) Windows is just a pain all around -- it stores the raid configuration in the registry (or used to).
The OS can generally reside anywhere as long as the core kernel knows how to mount it. In Linux and most other *NIX, that's a matter of keeping the kernel (plus boot loader and initrd) outside the array. For Windows, it's a little harder as various drivers have to get loaded.
[where array is something other than RAID1. RAID1 is just a bunch of disk clones.]
An ok "budget" fileserver with 1TB of space can be built for around 1k$. And maybe even cheaper if you don't care what it looks like:-)
I built a 1.4TB array for about 2500$ two+ years ago. That was with 80GB drives. So, I know damn well it can be done much cheaper than this guys box. (We tested it as a full news server for a month. [click] yeah, you lose a lot putting a filesystem on it.)
Why is it that Sun's drive model numbers are also specific to a firmware revision?...
Because Sun puts Sun specific firmware on the drives they supply. No one has ever told me exactly why Sun has to have their own firmware -- or Sun isn't telling anyone what parts of the specs they rewrote specifically to work with their own OSes. Arrays are sold with a set of TESTED drives. Every single drive has been spun-up and stress tested before it's shipped with the array. That's why they cost more than what you can get at the local 7-11.
Let me bring up up to speed on MTBF... it's just a f***in' number. It doesn't mean anything. And it means even less for IDE drives. (IDE drives are tested in batch... make 1000 drives, test 10 of them.)
Indeed. In all of my "servers", they have the absolute cheapest crap-video card I can find simply to make the BIOS shutup -- which happens to be a 6$ video seven AGP video card:-) (The server at my feet has an S3 Savage 4 litterally pulled from the trash...)
(And the "appliances" I build at work have the video hardware -- ATI Rage XL -- disabled.)
Bullshit. I'm certain no one has ever asked for their router to randomly redirect an http session for a "Parental Controls" feature. What people wanted was the PC feature, not a router that interferes with network traffic.
Now, if it were the default behaviour following the firmware update to redirect *ALL* http sessions until the feature is configured (yes/no/demo), then this would be acceptable. Stealing one connection seamingly at random is broken behavior for any network device.
Rest assured, I will not be buying Belkin shit. (Not even cables.) [Not that I have been, anyway.]
What if a backdoor was installed last week, or last month, but was not caught?
It would make no difference. The official kernel releases from Linus are exports of the BK tree. The BK->CVS gateway is one way. Changes to the CVS data are NEVER pushed back into the BK tree which is how this was detected -- any change to the CVS repo is automatically suspect.
The only non-x86 (ia32 and ia64) supported systems are IBM z, i, p, and S/390 systems. And then, only at the highest ("Advanced Server") version. What once could be obtained for $49.95 at Best Buy (the boxed redhat set) is now $179 -- without any technical support; it's $299 if you want any support.
small companies with 50 computers are lucky to get 8 IP addresses
And even that is generally more than they need. Those 50 computers are mostly people's desktops that certainly do not have any need for a global IP address all to itself.
advertising multiple IP blocks, bloating the internet's routing tables
This is simply not true. ISP's announce their netblock not the individual assignments within it. Anything smaller than a/20 is not guaranteed to be globally routed. Anything smaller that/24 is almost certainly not globally routed.
I've never looked at Progeny or Corel. I think Corel was what FatPipe used in their stuff after dumping Windows (NT?). Maybe that was Caldera. Anyway, it looked like a colorized version of the HP boot loader.
"Here, continue development for us so we can recoop our development costs." RedHat used to take all the code from other developers, futz with it and release it under their banner. Now, they are going to continue selling it; they just don't want to incur the cost of the "futzing" phase. On some level, that's a Good Thing (tm) as redhat has dug their fingers into everything they've shipped for the last few years. On all other levels, this may be the end of redhat as a common linux distro. (SLS anyone?)
Slashdot Mirror
Electricity was a luxury for a long time. Back then, people had been without it all their lives and thus used nothing that depended on it -- wood stoves, oil lanterns, horses and oxen, ice boxes, etc. Fast forward to today and everything runs on electricity. I'd bet a fair number of homes in the US don't have the materials to even start a fire without electricity. Shit, even lighters employ electricity -- piezo element striker (there's nothing to ever wear out like a flint.)
:-))
Just like everything else, once everyone has a thing for long enough, they don't know how to do without it. I grewup (well half way +/-) in a world without the computer -- in the 70's computers where the size of builds. I've worked with computers for ~20 years now; I wouldn't know how to function without them today.
HOWEVER, "broadband" is still a luxury and will remain so for many years to come. 56k modems (53k/48k/whatever) are perfect usable, but time consumingly slow. 128K ISDN/144K IDSL is quite sufficient for what 90% of the world does (legally) -- email, web browsing, IM, any number of games, etc. I used 28.8 dialup for ~5 years, and ISDN for ~7 years. I switched to an Earthlink cable modem about a month ago because it's 1/5th the cost -- yes, cost not bps was the reason and earthlink because they provide dialup for free when I'm out of town. (the increase in speed is a nice bonus, tho'
[1981 first introduction to a computer -- Tandy TRS Model I, 1984 first computer of my own -- Tandy Color Computer 2.]
[cramer:ttyp0]dominion:~/[10:43pm]:cat /etc/redhat-release
release 4.1 (Vanderbilt)
*whistles innocently*
"can run" as in VMS will know what to do with the specific hardware. "Alpha", like "PC", does not refer to a single hardware design. There were systems designed to run VMS/Unix and systems designed to run NT. (which mostly boils down to SRM vs. AlphaBIOS.)
"can run" as in VMS will recognize the hardware and go about it's business. That's like installing a BeOS developer release on a Xeon. (that probablly would work. it just wouldn't see anything other than the processor.)
Tru64 panics attmepting to determine the hardware... when it reads the hwdb, it doesn't understand what the cpu board id is and gives up. It's labeled "DIGITAL Server 3000", but DEC renamed the hardware three times before Tru64 existed... According to linux, it's a Noritake/Primo -- which is the correct MB/cpu codename.
PS: I don't care about VMS. I thought it was odd for them to send me VMS when I requested a "hardware update" CD (which is simply an AlphaBIOS and SRM update CD.)
VIM - Vi IMproved 4.2 (1996 June 17, compiled Dec 17 1996 14:55:08)
/dev/null /dev/null
/home/cramer/bin/elvis[ 9:17pm]:elvis --version
...
[cramer:ttya1]dominion:~/[9:16pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 0 Nov 11 21:16 test
[cramer:ttya1]dominion:~/[9:16pm]:echo ":wq" | vim test >
Vim: Warning: Output is not to a terminal
Vim: Warning: Input is not from a terminal
[cramer:ttya1]dominion:~/[9:16pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 0 Nov 11 21:16 test
[cramer:ttya1]dominion:~/[9:17pm]:echo ":wq" | elvis test >
[cramer:ttya1]dominion:~/[9:17pm]:ls -li test
403865 -rw-r--r-- 1 cramer root 1 Nov 11 21:17 test
[cramer:ttya1]dominion:~/[9:17pm]:which elvis
[cramer:ttya1]dominion:~/
ELVIS 1.8pl3, by Steve Kirkendall (22 April 1994)
RCS magic behind the scenes. (login to the CVS server and mv the RCS file.)
Most text editors work on an in-memory copy of the file: open, read, close, edit, open, write, close, exit.
Another option is to use the filesystems journal to rollback changes.
Heh, I have a copy of VMS (or is it MVS?) DEC sent it to me with a bunch of updates for my alpha. Of course, I have no license keys for any of :-) (And I don't think that alpha can run VMS.)
For the free license, the repo itself would have to be public. Open Logging does not send the changes, just a summary of the files changed and comments. And as of a few weeks ago, you couldn't browse the openlogging tree anyway :-) So, openlogging isn't the problem.
And no one has pointed out the obvious: what do you do when the CVS server dies? CVS is not distributed. There is exactly ONE copy of the repo. You cannot restore the repo from a checked out copy as it only carries one version of the file(s). BK is distributed; one works from a clone of the repo that can be used to restore a lost server.
[Personally, I prefer clearcase for my SCM needs. But it's not distributed either. And there's no flavor of it that's even remotely free. Hmm, a home dir as a clearcase dynamic view??? That'd be tricky.]
Well, are they being "rampantly" copied today? HDTV broadcast stations have been around for many years now. (WRAL-HD was the first one to be licensed back in 1998. I've not heard Mr. Goodman rant for even one second about people recording what he's been broadcasting for the past *5* years.)
Answer: No.
Or more accurately, no more so than analog broadcasts.
NetApp has "certified firmware" not "their own"... There's a difference between saying "we like fw v 4.1.0.5" and having your own custom "only given to us, made specifically for us" firmware. Sun (and Apple of the old days) have custom firmware. In fact, the drive ID says it's a sun (or apple) drive. NetApp drives are still seagate/quantum/whomever drives. And they use firmware you could find in retail drives.
NetApp's quality control requires all the drives have the same proven stable firmware so there are no gotcha's down the line. (I have several dozen drives from netapp arrays, btw.) Which means none of their drives run the latest junk from the factory.
[Please note, it is very rare to update the firmware on a hard drive. (or it used to be.) And it's equally difficult to get the update out of the manufacturer.]
I'll look around for the pictures of it before it was in a case. And after. It was sold to a company that handles processing for insurance forms -- it's their fileserver. (We rolled it out to a Chrysler LeBaron, loaded it in the back seat, and they drove it back to NY. [I'm in NC.])
Message from syslogd@gir at Fri Nov 7 13:02:53 2003 ...
...
gir kernel: MCE: The hardware reports a non fatal, correctable incident occurred on CPU 0.
Message from syslogd@gir at Fri Nov 7 13:02:53 2003
gir kernel: Bank 1: 9400000000000151
*whistles innocently*
Because they are installing windows (xp home?) on it. I'd also recommend avoiding NTFS on removable media outside of a domain -- bad things happen when one NTFS volume ends up on a different "alien" system.
:-) Granted, I spent several years on a VT420 attached to a sparc5 clone and a 486dx50 w/o a video card (yeah, it beeped but still continued to boot.)
Rule #1: Servers don't have video hardware - period.
I never could get an answer as to why all the Sun Enterprise servers at Interpath had Creator 3d video hardware in them? Hello, that's a 2000$ waste. (and why do people install all of X11 on systems that will never X applications?) They were web and email servers.
Maybe I'm a purest.
Well, mirroring (RAID1) is pretty painless on any system. Arrays are simply clones. It's very easy to deal with situations where a clone has dropped out.
For RAID0, a failure means lost data. "Please drive through."
For RAID>1, software based solutions can be a serious pain in the ass. Ok, what's the commands to quiet the bus for drive removal? what's the commands for restoring a drive? What happens when I update the kernel/raid software/etc.? I've had major trouble with linux basically having to bring the system back to the workbench to get RAID reactivated. I've had minor problems with Solaris -- patches breaking things until I read the solution to a non-public bug report. (A public patch breaks things but only contract customers are told how to fix it.) Windows is just a pain all around -- it stores the raid configuration in the registry (or used to).
"install into" != "boot from"
The OS can generally reside anywhere as long as the core kernel knows how to mount it. In Linux and most other *NIX, that's a matter of keeping the kernel (plus boot loader and initrd) outside the array. For Windows, it's a little harder as various drivers have to get loaded.
[where array is something other than RAID1. RAID1 is just a bunch of disk clones.]
An ok "budget" fileserver with 1TB of space can be built for around 1k$. And maybe even cheaper if you don't care what it looks like :-)
I built a 1.4TB array for about 2500$ two+ years ago. That was with 80GB drives. So, I know damn well it can be done much cheaper than this guys box. (We tested it as a full news server for a month. [click] yeah, you lose a lot putting a filesystem on it.)
- Why is it that Sun's drive model numbers are also specific to a firmware revision?
...
Because Sun puts Sun specific firmware on the drives they supply. No one has ever told me exactly why Sun has to have their own firmware -- or Sun isn't telling anyone what parts of the specs they rewrote specifically to work with their own OSes. Arrays are sold with a set of TESTED drives. Every single drive has been spun-up and stress tested before it's shipped with the array. That's why they cost more than what you can get at the local 7-11.Let me bring up up to speed on MTBF... it's just a f***in' number. It doesn't mean anything. And it means even less for IDE drives. (IDE drives are tested in batch... make 1000 drives, test 10 of them.)
Indeed. In all of my "servers", they have the absolute cheapest crap-video card I can find simply to make the BIOS shutup -- which happens to be a 6$ video seven AGP video card :-) (The server at my feet has an S3 Savage 4 litterally pulled from the trash...)
(And the "appliances" I build at work have the video hardware -- ATI Rage XL -- disabled.)
- This was by popular demand.
Bullshit. I'm certain no one has ever asked for their router to randomly redirect an http session for a "Parental Controls" feature. What people wanted was the PC feature, not a router that interferes with network traffic.Now, if it were the default behaviour following the firmware update to redirect *ALL* http sessions until the feature is configured (yes/no/demo), then this would be acceptable. Stealing one connection seamingly at random is broken behavior for any network device.
Rest assured, I will not be buying Belkin shit. (Not even cables.) [Not that I have been, anyway.]
- What if a backdoor was installed last week, or last month, but was not caught?
It would make no difference. The official kernel releases from Linus are exports of the BK tree. The BK->CVS gateway is one way. Changes to the CVS data are NEVER pushed back into the BK tree which is how this was detected -- any change to the CVS repo is automatically suspect.See also:
Red Hat Enterprise Linux Support Options and Pricing
Product Comparison Chart
The only non-x86 (ia32 and ia64) supported systems are IBM z, i, p, and S/390 systems. And then, only at the highest ("Advanced Server") version. What once could be obtained for $49.95 at Best Buy (the boxed redhat set) is now $179 -- without any technical support; it's $299 if you want any support.
- small companies with 50 computers are lucky to get 8 IP addresses
And even that is generally more than they need. Those 50 computers are mostly people's desktops that certainly do not have any need for a global IP address all to itself.- advertising multiple IP blocks, bloating the internet's routing tables
This is simply not true. ISP's announce their netblock not the individual assignments within it. Anything smaller than aI've never looked at Progeny or Corel. I think Corel was what FatPipe used in their stuff after dumping Windows (NT?). Maybe that was Caldera. Anyway, it looked like a colorized version of the HP boot loader.
"Here, continue development for us so we can recoop our development costs." RedHat used to take all the code from other developers, futz with it and release it under their banner. Now, they are going to continue selling it; they just don't want to incur the cost of the "futzing" phase. On some level, that's a Good Thing (tm) as redhat has dug their fingers into everything they've shipped for the last few years. On all other levels, this may be the end of redhat as a common linux distro. (SLS anyone?)
ftp.redhat.net:/pub/redhat/linux/9/en/os/.... ... ... ... ../ .... ..
-> i386
ftp.redhat.net:/pub/redhat/linux/8.0/en/os/
-> i386
ftp.redhat.net:/pub/redhat/linux/7.3/en/os/
-> i386
ftp.redhat.net:/pub/redhat/linux/7.2/en/os/
-> i386, ia64, s390
ftp.redhat.net:/pub/redhat/linux/7.1/en/os/
-> alpha, i386, iSeries, ia64, pSeries, s390x
ftp.redhat.net:/pub/redhat/linux/7.0/en/os
-> alpha, i386
ftp.redhat.net:/pub/redhat/linux/6.2/en/os/
-> alpha, i386, sparc
...
v7.2 is circa Jun/Jul 2002. Yes, there are updates avialable, but no recent, updated release.
[Note: The link you provide is for RedHat ENTERPRISE Linux which will still be available to those willing to part with a lot of cash.]