I very much second this. I attended Operation Catapult before my Senior year of high school, and not only had a blast, but learned a ton that I took off to college with me.
OK, I'll bite, why would you -want- to stick with 2.4? 2.6 has been out for a couple years now, I think we can safely say it's useable.
Yes, there's something to be said for stability, and I won't argue against Debian's stability, but I've got hardware for which there is no support. I'm not going to run everything on 2004 hardware just because Debian insists on running 2004 software.
Look at some of the stranger RAID options. If you just use RAID5, you'll be selling yourself short. RAID3 is worth a look. I'd actually suggest you put two controllers in a machine. Run RAID0 on 4 drives on a single controller. Run RAID0 on 4 drives on the other controller. Then use Windows or Linux software RAID to run RAID1 between the two RAID0 drives. Very fast performance and fully fault tollerant.
Uhh. Yes. Then you can lose one disk in each side, and you have lost all your data.
This would perhaps be slightly less than fully fault tolerant.
Perhaps you meant to set up 4 mirror pairs, 2 on each controller, and use software to RAID0 them together.
I have successfully done this with a 24 disk 5U chassis, and it is an IO steamroller (our database server, right now).
Just to alleviate some of his bandwidth, I have mirrored the mail_filtering pages. Looks like it's all there. Let me know if you want me to take it down.
Money exists only to be spent. Technology exists mostly to be bought (and occasionally to just look cool on its own). Politicians are known for spending money like it's going out of style.
Ergo, let's go burn billions on this new technology that wouldn't have made a damn bit of difference anyway. Cuz, like, it's cool.
First to use BGP you must have a block of 16 C class's (a/20) either from ARIN ($2500 per year) or from one of your ISPs (variable rates)
This is the smallest block that the internets border routers will 'see' a BGP route for.
Anything smaller will be ignored and you will not be able to reach those networks.
Actually, this isn't really true. Depending on where you are in the IPv4 namespace, you can get announcements as small as a/24 through backbone filters. Now, if you get space under the Swamp (below 192.0.0.0), you'll probably need the full/20. Otherwise, well, I've had single/24's be reachable from everywhere (in 216.249.0.0/16).
Of course, you don't really need that much space. Get two T1's (one from each of two providers), a/24 from one (or both), and cross-announce over BGP. Voila. Redundant links. Maybe not very well load balanced (depends on who your two providers are, and where your traffic goes), but redundant.
The problem wasn't really MySQL at all, it was FreeBSD's threading support in 4.x. It's crap, to say the least, at least as far as a DBMS is concerned.
As the previous poster said, hopefully 5.0 will help.
But for the Open Source products this was not the case. Contacting the maintainers of the Open Source products and asking if anyone provided commercial support was fruitless; in one case the response was downright rude (basically a variation on RTFM) and in the other the response was more helpful, but still could not suggest anything other than being active on the mailing list.
Er, well. Ok, yes, it's kinda rude, but it's also a reasonable expectation that you've read the manual and tried basic troubleshooting ("Is the monitor plugged in, Mrs. Jones?"). Remember, support contracts exist to save your ass when you've broken it and, though you've tried, you can't fix it. Not to hold your hand as you plug in the power on the disk array, replace the fibre-channel fiber patch that security cut with a pair of hedge clippers, or replace a smoking powersupply.
Actually, I -have- seen the newer Google racks. They're still using Rackable Systems' 1- and 2-U servers, and they most certainly are -NOT- diskless.
The last time I saw them was when I was last in the San Jose MFNX datacenter (and, coincidentally, recognized all the servers that had been moved there from Exodus/GlobalCenter in Sunnyvale, where my former employer's equipment is still colocated). Still normal PCs with disks. Most of their servers are, as I said, the Rackable Systems' machines, with servers on both sides of the rolling racks, and an HP ProCurve switch in the middle on each side, running 100BaseTX to each box and Gigabit to the upstream.
Yeah, the lady at Chili's said you left about 3 minutes before we got there.
PopeFelix (the host) had to cancel at the last minute because of work, so myself, kimi, and Hlynna showed up at about 7:15. We waited until 7:45, then got a table and had a few drinks, then left.
I wore my first-edition green-and-white Slashdot hat, hoping people coming for the meet would get the picture. But I guess we were there after you left.
If I put a bunch of fans on a case that are blowing air out (better wording than "sucking air out"?), and dont have sufficient ventilation holes elsewhere in the case to allow air to come in, the resultant low-pressure inside the case will strain the fan. It's trying to move air that doesn't WANT to move (remember, a fluid will go from a high pressure area to a low pressure area freely, but you have to force the reverse).
(Or, as you said...
If anything might damage the fans in this situation, it's the greater external air pressure pushing against them.
That's what I meant.)
If I put some on one side blowing in, and some on the opposite side blowing out, I've created a "duct" of air that will rapidly flow through and cool the system. This is the ideal.
And yes, as another poster pointed out, a fan would break down in a vacuum, because of the lack of heat dissipation (but not because of being overworked).
3) Add as many fans as you can...and point them all blowing out of the case to try and reduce the pressure inside.
Actually, what you want is to point some of them out, and some of them -in-, and create a constant flow of air through the system.
Make sure there are -lots- of air vents so you can move monster amounts of air through the system, and make sure you have equal capacity of fans sucking air in as you do blowing air out. "Depressurizing" the system (to the minor extent that your average case fan will do, anyway) is bad. Heat exchange requires something to dissipate heat into. If you thin the air, you dissipate less heat.
If all you do is try to "depressurize" the system, you may not only damage the CPU by not cooling it sufficiently, but you might also put an unnecessary strain on your cooling fans, that are constantly laboring to suck air out.
Yeah, I started in 96, as a CS student. Unfortunately, I never graduated. In March of 1999, I was in a near-fatal auto accident, two days before I was to return to class from coop. I was in the hospital here at home in South Carolina for a month, and in a wheelchair for two more. Needless to say, Drexel didn't see this as a valid reason to miss class, and killed all my financial aid. So, I've not returned.
This is the kind of thing that -will- have to be planned, and to get the kind of support you want, you should enlist the help of others in your organization.
First step. Identify what people use now. Aside from the obvious Windows version $foo, you've probably got IE, Office, etc. Also identify other non-Microsoft applications that people use that still require Windows.
After you've figured out what apps you have to replace, start looking for the open source replacements. Obviously for IE, you've got Opera, Netscape, Mozilla, Konquerer, etc. Grab them all, put them on a machine. Bring in some of your employees, have them test out browsers. See which one works best for them (or, give them the option). Grab StarOffice or OpenOffice, and sit them down, show them how it works. Get their feedback. If SO/OO doesnt cut it, look for other options (KOffice, AbiWord, WordPerfect, etc). Do this for all the apps you can find Linux replacements for.
For the programs you can't replace, look into compatibility with WINE. I'd be willing to bet there are Win32 accounting applications in use that probably can't be replaced for free with a Linux version (however, since I haven't looked lately, I very well could be wrong...). So, your next best bet is to see if they work with WINE. That may be the only way to get them working.
You probably won't be able to replace every app in use in your organization with a free Linux workalike. Some may have to be grandfathered in WINE for a while. And you may not be able to replace every Windows install with a Linux install. But if you do your planning properly, keep other people informed and take their feedback, I'm sure you can get the support you need to push for a near-total migration. And when you do, don't stop looking. New apps come out all the time. Something you can't replace now may be replaceable in 3 months.
Slashdot Mirror
Two trillion searches per year is ~64,000 searches per second.
That number is very not impressive.
I suspect the correct number is at least one, if not two, orders of magnitude higher.
That's funny, my employers have always expected me to.
I very much second this. I attended Operation Catapult before my Senior year of high school, and not only had a blast, but learned a ton that I took off to college with me.
I am a Senior Service Engineer at Yahoo!.
I work on Y! Buzz (buzz.yahoo.com), the new Y! Front Page (m.www.yahoo.com), and My Yahoo! (my.yahoo.com).
I don't have a degree.
It's not just "announced". It's available for sale.
OK, I'll bite, why would you -want- to stick with 2.4? 2.6 has been out for a couple years now, I think we can safely say it's useable.
Yes, there's something to be said for stability, and I won't argue against Debian's stability, but I've got hardware for which there is no support. I'm not going to run everything on 2004 hardware just because Debian insists on running 2004 software.
Look at some of the stranger RAID options. If you just use RAID5, you'll be selling yourself short. RAID3 is worth a look. I'd actually suggest you put two controllers in a machine. Run RAID0 on 4 drives on a single controller. Run RAID0 on 4 drives on the other controller. Then use Windows or Linux software RAID to run RAID1 between the two RAID0 drives. Very fast performance and fully fault tollerant.
Uhh. Yes. Then you can lose one disk in each side, and you have lost all your data.
This would perhaps be slightly less than fully fault tolerant.
Perhaps you meant to set up 4 mirror pairs, 2 on each controller, and use software to RAID0 them together.
I have successfully done this with a 24 disk 5U chassis, and it is an IO steamroller (our database server, right now).
Ha ha. Very funny.
Just to alleviate some of his bandwidth, I have mirrored the mail_filtering pages. Looks like it's all there. Let me know if you want me to take it down.
emerge sync && emerge -uD world
[go get coffee, depending on how long it's been, it could be a while]
etc-update
[done!]
Money exists only to be spent. Technology exists mostly to be bought (and occasionally to just look cool on its own). Politicians are known for spending money like it's going out of style.
Ergo, let's go burn billions on this new technology that wouldn't have made a damn bit of difference anyway. Cuz, like, it's cool.
Yes, but Yahoo! and Google have people who develop for FreeBSD and Linux, which qualifies as "support".
I don't write FreeBSD kernel drivers or Linux load balancing code, so I wouldn't qualify as support for my company.
Actually, this isn't really true. Depending on where you are in the IPv4 namespace, you can get announcements as small as a /24 through backbone filters. Now, if you get space under the Swamp (below 192.0.0.0), you'll probably need the full /20. Otherwise, well, I've had single /24's be reachable from everywhere (in 216.249.0.0/16).
Of course, you don't really need that much space. Get two T1's (one from each of two providers), a /24 from one (or both), and cross-announce over BGP. Voila. Redundant links. Maybe not very well load balanced (depends on who your two providers are, and where your traffic goes), but redundant.
The problem wasn't really MySQL at all, it was FreeBSD's threading support in 4.x. It's crap, to say the least, at least as far as a DBMS is concerned.
As the previous poster said, hopefully 5.0 will help.
Er, well. Ok, yes, it's kinda rude, but it's also a reasonable expectation that you've read the manual and tried basic troubleshooting ("Is the monitor plugged in, Mrs. Jones?"). Remember, support contracts exist to save your ass when you've broken it and, though you've tried, you can't fix it. Not to hold your hand as you plug in the power on the disk array, replace the fibre-channel fiber patch that security cut with a pair of hedge clippers, or replace a smoking powersupply.
Actually, I -have- seen the newer Google racks. They're still using Rackable Systems' 1- and 2-U servers, and they most certainly are -NOT- diskless.
The last time I saw them was when I was last in the San Jose MFNX datacenter (and, coincidentally, recognized all the servers that had been moved there from Exodus/GlobalCenter in Sunnyvale, where my former employer's equipment is still colocated). Still normal PCs with disks. Most of their servers are, as I said, the Rackable Systems' machines, with servers on both sides of the rolling racks, and an HP ProCurve switch in the middle on each side, running 100BaseTX to each box and Gigabit to the upstream.
Yeah, but the latency's a bitch.
Yeah, the lady at Chili's said you left about 3 minutes before we got there.
PopeFelix (the host) had to cancel at the last minute because of work, so myself, kimi, and Hlynna showed up at about 7:15. We waited until 7:45, then got a table and had a few drinks, then left.
I wore my first-edition green-and-white Slashdot hat, hoping people coming for the meet would get the picture. But I guess we were there after you left.
Hopefully next month.
You misunderstand what I mean.
If I put a bunch of fans on a case that are blowing air out (better wording than "sucking air out"?), and dont have sufficient ventilation holes elsewhere in the case to allow air to come in, the resultant low-pressure inside the case will strain the fan. It's trying to move air that doesn't WANT to move (remember, a fluid will go from a high pressure area to a low pressure area freely, but you have to force the reverse).
(Or, as you said...
If anything might damage the fans in this situation, it's the greater external air pressure pushing against them.
That's what I meant.)
If I put some on one side blowing in, and some on the opposite side blowing out, I've created a "duct" of air that will rapidly flow through and cool the system. This is the ideal.
And yes, as another poster pointed out, a fan would break down in a vacuum, because of the lack of heat dissipation (but not because of being overworked).
3) Add as many fans as you can...and point them all blowing out of the case to try and reduce the pressure inside.
Actually, what you want is to point some of them out, and some of them -in-, and create a constant flow of air through the system.
Make sure there are -lots- of air vents so you can move monster amounts of air through the system, and make sure you have equal capacity of fans sucking air in as you do blowing air out. "Depressurizing" the system (to the minor extent that your average case fan will do, anyway) is bad. Heat exchange requires something to dissipate heat into. If you thin the air, you dissipate less heat.
If all you do is try to "depressurize" the system, you may not only damage the CPU by not cooling it sufficiently, but you might also put an unnecessary strain on your cooling fans, that are constantly laboring to suck air out.
Actually, the specific 800lb gorilla I was thinking of was a client when Steve and I were working together.
eBay.
They -are- geographically distributed.
I've seen one of Google's colo cages. It truly is badass.
Actually, I believe there are around 8,000 CCIE's now. Not sure of the exact number, but I'm sure if you asked Cisco, you could get it.
Yeah, I started in 96, as a CS student. Unfortunately, I never graduated. In March of 1999, I was in a near-fatal auto accident, two days before I was to return to class from coop. I was in the hospital here at home in South Carolina for a month, and in a wheelchair for two more. Needless to say, Drexel didn't see this as a valid reason to miss class, and killed all my financial aid. So, I've not returned.
:)
Yet all my accounts still work
This is the kind of thing that -will- have to be planned, and to get the kind of support you want, you should enlist the help of others in your organization.
First step. Identify what people use now. Aside from the obvious Windows version $foo, you've probably got IE, Office, etc. Also identify other non-Microsoft applications that people use that still require Windows.
After you've figured out what apps you have to replace, start looking for the open source replacements. Obviously for IE, you've got Opera, Netscape, Mozilla, Konquerer, etc. Grab them all, put them on a machine. Bring in some of your employees, have them test out browsers. See which one works best for them (or, give them the option). Grab StarOffice or OpenOffice, and sit them down, show them how it works. Get their feedback. If SO/OO doesnt cut it, look for other options (KOffice, AbiWord, WordPerfect, etc). Do this for all the apps you can find Linux replacements for.
For the programs you can't replace, look into compatibility with WINE. I'd be willing to bet there are Win32 accounting applications in use that probably can't be replaced for free with a Linux version (however, since I haven't looked lately, I very well could be wrong...). So, your next best bet is to see if they work with WINE. That may be the only way to get them working.
You probably won't be able to replace every app in use in your organization with a free Linux workalike. Some may have to be grandfathered in WINE for a while. And you may not be able to replace every Windows install with a Linux install. But if you do your planning properly, keep other people informed and take their feedback, I'm sure you can get the support you need to push for a near-total migration. And when you do, don't stop looking. New apps come out all the time. Something you can't replace now may be replaceable in 3 months.