Yes, this is exactly what I was thinking. The advantage is obvious; you can get by with very cheap hardware, failover time is in the order of minutes (although manual intervention is required). The intermediate versions are also useful, I'd imagine; might be worth checking up on someone who use rsync for backups with hard links which did a similar job.
One thing to add; I really hope you're not relying on the backup machine as your sole source of backups; if you lose the site (fire/flood), you lose all your data. At a minimum, take weekly backups offsite (taking tapes home is acceptable, although make sure tapes are secured), ideally daily.
Finally, you shouldn't need to reconfigure clients; just make the backup server appear to be the main server by changing its IP address and netbios name (if using Samba).
I'm with grigori; never had a problem with power failures on a properly configured Sun box. The only server I know of which has a problem is the E150, which you can't power on from init 5 without either a keyboard or a screwdriver. The E150 is probably the worst put together piece of hardware I've ever seen and it astounds me that Sun released it.
Besides, power failures shouldn't happen; you should have UPS on all important servers so power failures shouldn't be a problem at all.
Well, there are other issues like lock contentions (making sure 2 CPUs aren't using the same chunk of RAM) but the point is correct. The problem with 16 500MHz cores is that a single threaded app will still run at the same speed as a single 500MHz core; you would be able to run more of them at the same time on the 16 cores, though.
Hah, I take your $30 10' x-over cable and raise it with a 98 UKP crossover cable (10 feet long). Required two of them so that our Sun Cluster would be supported, go figure.
Possibly; think about how much a mobile disrupts a telephone conversation (I'm sure you've heard the bleeping on the phone while your mobile receives a text message!). When I first got a mobile (about 7 years ago), I could tell it was going to ring before it made a noise, simply because of the disruption on the monitor.
As for the real impact, well, I work in an office where almost everyone has a mobile and the computers are generally well-behaved.
Closer to the truth than you might like... the legislators have their backers who will buy their own clauses in the statutes and the lawyers will argue for whoever is paying their bills.
Who will define it? The lawmakers and their sponsors. Once the law is in place, lawyers and judges will have their pop at anything which isn't 100% crystal clear.
Re: training: on any major upgrade of windows, you usually need some kind of training anyway, so the assumption that "switching to linux means retraining" is an exaggeration.
My guess is that it's a fairly specific, non-standard load that will garner a 1000x gain. Think about how this would help 10,000 processes accessing 10,000 different files and reading in their contents; under a normal system, each process would get a short time to read data, then the next process etc. With the read-ahead, it's likely that the kernel would read in larger chunks of data at a time and cut down on disk seeks.
Check the display drivers if you're in a Compaq shop; ours have a little trick where Ctrl-Alt-Down Arrow inverts the display and Ctrl-Alt-Up Arrow reverts it to the right way round.
2*physical RAM is an old rule of thumb which falls over on large systems; we have one server with 16GB of RAM and we don't really want to allocate 32GB of swap space which should never be used as a rule.
In most cases, servers should never swap; most systems these days only have swap as a space for crash dumps; both Solaris and Windows do this.
Also, as others have said, leave Windows page files at a constant side to avoid fragmentation; this is one thing which the *nix world has definately got right.
Part of what I had to learn in primary was my times tables; we'd have to memorize everything from 1*1 up to 12*12 (and all the numbers in between). It was very boring and I hated it at the time, but I'm glad of it now, as I can multiply pretty well in my head.
In a coincidence, the Sun advert about Java being in the Spirit Rover was showing at the top of the front page (and on this page as I write this post).
They're not actually trying to patent virtual desktops, they're trying to patent a pager with a preview of each desktop. You know, kind of like Gnome has (and probably KDE as well; can't remember).
Doesn't make it that much better, but at least make sure you're ranting about the right thing.
I think it's on install CD 2, but it's installed on a full installation (SUNWCall). I also think it was added in Solaris 8, so 2.6 or 7 didn't have it. Solaris 8 also added gzip (thankfully) and a few other open source packages.
BTW, all packages prefixed by SUNW are Sun packages.
Mice don't always need a flat surface; I use a Microsoft Intellieye mouse, one of the optical ones. It works fine using my leg as a surface to track on.
"killall daemon" works just as well - in fact, that's all most stop scripts do
Just make sure you don't transfer your script to Solaris, or unexpected results will occur. Hint: under Solaris, killall does exactly what it says on the tin; it kills all processes. Since Solaris 7, it's provided pkill for the same (if not greater) functionality as killall under linux etc.
Monitoring/limiting bandwidth usage by clients at the ISP end probably incurs a performance penalty while you're monitoring; therefore it's probably cheaper to have the client check it.
Seriously, check out NTK which occassionally features "one of these pictures is not like the rest" links to google images searches, where normally innocent searches pull up some "interesting" results.
Slashdot Mirror
One thing to add; I really hope you're not relying on the backup machine as your sole source of backups; if you lose the site (fire/flood), you lose all your data. At a minimum, take weekly backups offsite (taking tapes home is acceptable, although make sure tapes are secured), ideally daily.
Finally, you shouldn't need to reconfigure clients; just make the backup server appear to be the main server by changing its IP address and netbios name (if using Samba).
Besides, power failures shouldn't happen; you should have UPS on all important servers so power failures shouldn't be a problem at all.
Well, there are other issues like lock contentions (making sure 2 CPUs aren't using the same chunk of RAM) but the point is correct. The problem with 16 500MHz cores is that a single threaded app will still run at the same speed as a single 500MHz core; you would be able to run more of them at the same time on the 16 cores, though.
Hah, I take your $30 10' x-over cable and raise it with a 98 UKP crossover cable (10 feet long). Required two of them so that our Sun Cluster would be supported, go figure.
As for the real impact, well, I work in an office where almost everyone has a mobile and the computers are generally well-behaved.
Closer to the truth than you might like... the legislators have their backers who will buy their own clauses in the statutes and the lawyers will argue for whoever is paying their bills.
Who will define it? The lawmakers and their sponsors. Once the law is in place, lawyers and judges will have their pop at anything which isn't 100% crystal clear.
Re: training: on any major upgrade of windows, you usually need some kind of training anyway, so the assumption that "switching to linux means retraining" is an exaggeration.
I think we can safely guarantee the condition of just about any cargo which hits the moon at that speed...
My guess is that it's a fairly specific, non-standard load that will garner a 1000x gain. Think about how this would help 10,000 processes accessing 10,000 different files and reading in their contents; under a normal system, each process would get a short time to read data, then the next process etc. With the read-ahead, it's likely that the kernel would read in larger chunks of data at a time and cut down on disk seeks.
Check the display drivers if you're in a Compaq shop; ours have a little trick where Ctrl-Alt-Down Arrow inverts the display and Ctrl-Alt-Up Arrow reverts it to the right way round.
In most cases, servers should never swap; most systems these days only have swap as a space for crash dumps; both Solaris and Windows do this.
Also, as others have said, leave Windows page files at a constant side to avoid fragmentation; this is one thing which the *nix world has definately got right.
Part of what I had to learn in primary was my times tables; we'd have to memorize everything from 1*1 up to 12*12 (and all the numbers in between). It was very boring and I hated it at the time, but I'm glad of it now, as I can multiply pretty well in my head.
I do, but I read Slashdot too much and keep running out of free pages :)
As the advert says, find out at Sun.com/mars
Since when has that stopped anyone?
Shouldn't need to be; most of that should be handed off to the PAM modules.
Doesn't make it that much better, but at least make sure you're ranting about the right thing.
I'm not sure what the worst option is.
(before anyone says, yes, I'm aware that the article doesn't actually mention an alliance, but this was too good to pass up :) )
BTW, all packages prefixed by SUNW are Sun packages.
# pkginfo SUNWbash
system SUNWbash GNU Bourne-Again shell (bash)
Perhaps not always installed by default, but it is available. That's on Solaris 8, BTW. As for other stuff, check out www.sunfreeware.com
Mice don't always need a flat surface; I use a Microsoft Intellieye mouse, one of the optical ones. It works fine using my leg as a surface to track on.
Monitoring/limiting bandwidth usage by clients at the ISP end probably incurs a performance penalty while you're monitoring; therefore it's probably cheaper to have the client check it.
Seriously, check out NTK which occassionally features "one of these pictures is not like the rest" links to google images searches, where normally innocent searches pull up some "interesting" results.