Seriously- If you haven't enabled MFA on your Gmail account then please don't complain when you get hacked. It takes a couple of minutes- you have no excuse not to.
Yes- because they used CO2. CO2 forms carbonic acid in the blood stream which upsets the acid-base homeostasis. Our brains detect that imbalance which is what triggers the breathing reflex. The more CO2 you give them- the higher the level of carbonic acid- and the stronger the reaction. In other words- it's not lack of oxygen that causes us to want to breathe- it's too much CO2.
If you fill a room with only Nitrogen, or just place a non-rebreather mask on their face and connected to a Nitrogen tank- they will simply fall unconscious and die within a couple of minutes. There is no pain, and no panic.
I've used FreeBSD for as long as I can remember. My desktop is _still_ FreeBSD.
Having said that- nothing performs under high load situations like Solaris. I was not happy with Solaris in the 7, 8 or 9 days (Mostly because they kept changing things- seemingly at random), but Solaris 10, and especially OpenSolaris, have been rock solid and cool as can be under high load.
My current OpenSolaris heads are running on Dell 2950's with 32 GB of RAM, Myricom Dual Port 10Gb ethernet nics, 8 lane (2 x4) 3Gb SAS controller, dual port LP10000exDC Emulex 2Gb/s Fiber Channel cards, and dual 3GHz quad core processors. They're connected to 6 Promise vTrak J610s Dual controller disk shelves. The shelves have 4 32GB write optimized SSD's in a mirrored/striped config spread across both controller channels for our ZIL. There are 80 15k SAS drives and 8 256GB Read optimized SSD's as the L2ARC.
Both heads are connected to the shelves in an active/passive configuration in which the second head can immediately pick up if the first one fails.
We use COMSTAR to export ZFS volumes via iSCSI (over 10Gb ethernet to our VMWare servers) and Fiber Channel (To our older Sun QA Farm).
You can scale ZFS pretty amazingly large- you just have to architect the system.
We've got 80 15k SAS spindles spread across two 4 lane 3Gb/s (12 Gb/s x 2) controllers. We've got 4 32GB write optimized SSD's in a mirror+stripe configuration, two disks on each channel for our ZIL. We've got 32 GB of RAM for dedupe and our L1-ARC. We've got 8 read optimized 256GB SSD's as our L2-ARC read cache. There are four hot spares in our system- 2 300GB and 2 600GB to match the two sizes of disks that we have.
The system absolutely flies. We've got dual 10Gb ethernet links in each of the two server heads. The heads are active/passive and both are connected to all the disks at all times via active/active SAS switches in the disk shelves (Promise vTrak J610S Dual's)
We use COMSTAR for iSCSI and FiberChannel target capabilities. (The fiber channel is 2Gb for our old Sun QA servers.
All of this still came out to less than 1/4 of what NetApp wanted for a similar system.
Routing and ISP's are huge topics- what are you trying to do?
The main problem with routing isn't bandwidth- anyone can pump enough 1500 or 9000 byte frames per second to fill a gigabit pipe. The problem is when you have lots of small packets. At that point, dedicated routing hardware with a high-speed TCAM becomes really important.
What kind of line cards do you need? ADSL? Ethernet? OC12?
What kind of services do you need to run? BGP? OSPF?
What kind of bandwidth are you going to be pushing?
The basic Brother 4040CDN I just bought is a color laser with 4 individual toner cartridges, duplex, 21ppm and networking for $360- and it will last me a decade at least. It's not a photo printer, but if all you need is color it's a better choice than an overpriced inkjet.
Funny- but lots of houses in Levittown had copper radiant floor systems that failed and were attributed to the copper. We also have houses in Florida and Long Island developing pinhole leaks- either due to impurities in the copper, or environmental degradation. Either way- a single, jointless run of PEX is a hell of a lot safer bet (especially considering it can flex) than a copper pipe that can corrode, has lots of joints, and can't flex.
This is actually a law in a lot of states. The actual wording varies from "Left lane is for passing only" to "you can be in the left lane but must not impede faster traffic" to "you can be in the left lane, but must be driving at the speed limit and should not impede faster traffic."
Stop worrying about what's going to make you the most money and figure out what you enjoy. An MBA that hates his job is worthless. A computer scientist that isn't passionate about math and theory is worthless. An IT guy that isn't obsessed with all things tech will never be as good as the guy that is.
Figure out what you love doing and do that. If you really love it you'll be better at it. The best people in any field always make plenty of money.
As an aside- the last thing this world needs is more lawyers. The second to last thing this world needs is more MBA's.
Volkswagen has been making a diesel Golf that gets 60+ miles per gallon for several years now. The only "news" here is that the European division of Ford isn't as incompetent as their American division.
Any woman that is more concerned with electing a woman to the presidency, than they are with electing someone who represents their values, should lose their right to vote. It's idiotic in the extreme. If your grandmother wants to elect a fossil to the presidency because she's old, hey be that stupid. If she wants to elect McCain because he nominated a woman for the vice presidency, then your grandmother is an idiot who is playing right into McCains plans.
He nominated Palin because he wanted to tap the Hillary vote. If you think women are too stupid to see past such a ruse- then either you're also an idiot, or this country is hopelessly lost.
Did you even bother to read the original post? The person said their client wanted read-only access. The problem isn't the client messing up data- it's the client bringing the server to its knees with an abysmally inefficient query. Oracle has some features to limit the damage a user could do- but the only truly safe option is a read-only replica.
The person also wrote that the customer did not want 24 hour old data (you really didn't bother to read the post did you?) so your cube is a useless idea.
If you think it's a good idea to give clients direct access to your production database then please send me over your resume so I can make sure it goes on our "Never, ever hire this person" board.
They are incompetent- completely and utterly incompetent. They know only what encase or another piece of forensic software tells them. If the disk blocks have been rewritten a couple of times- they're not going to find it. They're not going to break AES unless you've done something stupid and left the key laying around.
The real bitch of it is- these guys never get challenged properly- especially in child porn cases. (Thank John Walsh- Adam's Law is absurd). They can claim whatever they want and the defense is basically helpless. The defense is not allowed to have their own copy of the drive to do forensic analysis on. They have to do it at the FBI lab with FBI equipment and with FBI goons hanging over their shoulders. If the FBI finds "overwritten" evidence- there is no good way to challenge that. It's your word against theirs.
Chain of custody? HAH! I've watched these guys leave crime scenes with drives under their arms, I've watched them run programs and click around a system they suspect of containing illegal material. No effort made to prevent trojans or other programs from covering their tracks. No effort made to preserve the state of the system. It's laughable.
And no- I wasn't a target. I did "forensic" analysis for years and got sick of watching these people make a mockery of my profession. (I put forensic in quotes because there is nothing scientific about these analyses- they are the best guesses of someone who may or may not be even remotely qualified to give an opinion).
What the hell do fingerprints have to do with this? You can't find a persons relatives via a finger print. You can't say "hey- think fingerprints looks almost like this other guys, so it much be a relatives!"
With DNA, you're using a DNA sample from a crime scene and matching it to a known criminals DNA to find a relative.
What you're suggesting is using a fingerprint from a crime scene, matching it to a known criminal, and then using that to find the persons relatives. That doesn't make the slightest bit of sense. If the fingerprint matches, you know your criminal. If it doesn't, you've got to keep investigating. Who they're related to isn't exactly important.
In this case we're talking about casting suspicion over people simply because their DNA is close to someone else's- that's frightening.
I love postgres- use it to handle millions of queries per day- but let's not kid ourselves: postgres replication in the form of slony blows. I have a master DB and it plublishes to a dozen read only databases. Managing that with slony just plain sucks. The simple fact is that setting up replication with mysql is dirt simple and that's part of why people use it.
I'm thinking of giving EnterpriseDB and their custom replication engine a try.
Those are the questions we can't get people to answer now. There are plenty of harder questions we can ask, but we'd be happy if we could find people who can answer those.
If you put 5 years of netwroking experience we're going to ask you about the 5 types of OSPF LSA's and what a "Not So Stubby Area" is. If you admin a website we're going to ask you what the primary differences between HTTP 1.0 and 1.1 are. If you put security we're going to expect you to explain public key cryptography in some reasonable detail. If you put programming down you damned sure better be able to explain NP completeness and know how to write a compiler. There are tons of questions to ask, including plenty of trick questions, but like I said- I'd be happy to find someone who can answer the basics at this point.
Does DNS use TCP or UDP? Explain. (The correct answer is not "TCP is used only for zone transfers.")
What port does service XYZ run on where XYZ is anything from SMTP or SSHD to NFS. (It's not about the right answers- if someone says I'd have to check/etc/services, that's a good answer too.)
What information is stored in an inode?
Explain the boot process, from the BIOS on up to a fully running linux kernel.
What is an initrd and why does it exist?
You have a system with ABC problem- how would you troubleshoot it?
How would you stop a RedHat system (or any other) with two network interfaces on two different networks from routing traffic between those interfaces? Do this on a running system, and do it on startup. There are several ways. Name as many as you can.
What is a file's link count and what does it mean?
What does the link count mean with respect to a directory? How many links would a particular directory have? Why?
I work for a large Unix (primarily Linux) shop and we've been interviewing people for a year now. We're in NYC so there should be plenty of candidates- and there are. 1 in 10 passes our basic screening test (10 questions and not difficult). We then interview the ones we think are competent in person. It's a tough interview- but you don't need to get everything right- just handle yourself well and show some thought. Hell I'd settle for someone saying "I'd need to read the man page in response ot every question we ask).
Only two candidates have passed, and both of those only barely. We considered them entry level people and were willing to pay them entry level rates ($70k - $80k) but they wanted 6 figures. These are people that would have required significant training to reach a level we consider good.
Maybe we just have high standards but we haven't seen a single serious candidate in the last year. Say what you want but the market is definitely tight- at least for a serious Unix admin.
The one thing that has to stop is people lying about their experience. Do not list "networking experience including BGP" if all you've ever done is type "router bgp blah" and "network blah." If you don't know how OSPF works and can give me a decent explanation, then don't list OSPF. Don't list X years of RedHat experience and RedHat certification if you can't tell me how to find out what version of RedHat a box is running. Don't tell me Linux experience if you're not familiar with basic troubleshooting tools like iostat and vmstat. Don't list Linux if you can't tell me something as simple as how to turn off routing in your distro of choice. Or if you don't know- at least don't suggest stupid things like disabling one of the two interfaces or adding (or removing) a route. It's really annoying to hear such stupid answers.
I second the Voicepulse Connect recommendation. Their web page sucks, a lot of information is missing, but in the end they're not doing that to hit you with hidden fees, their web department just looks to be incompetent:)
JHust to clarify- the company mentions possibly going to 28 stacked chips which would be 448 gigabits (not gigabytes) of storage- or about 56 GB of space. Now as flash chips grow in size- this could double (assuming 32 Gb NAND chips which are becoming available) to 96 or 112 GB of storage or more (assuming larger chips).
Slashdot Mirror
https://www.tomsguide.com/us/iphone-8-benchmarks-fastest-phone,review-4676.html
Took .7 seconds to open a PDF on the iPhone 8 versus 6.5 seconds on the Galaxy Note 8.
iPhone 8 had twice the graphics performance of the Galaxy Note 8. Like to play games on your phone- CPU and GPU matter.
So the various independent sites that have tested the chips and confirmed the performance are in on the conspiracy?
https://www.pcper.com/reviews/Processors/Apple-A11-Performance-Review-iPhone-8-Plus-Taking-Desktop
https://www.tomsguide.com/us/iphone-8-benchmarks-fastest-phone,review-4676.html
Seriously- If you haven't enabled MFA on your Gmail account then please don't complain when you get hacked. It takes a couple of minutes- you have no excuse not to.
Yes- because they used CO2. CO2 forms carbonic acid in the blood stream which upsets the acid-base homeostasis. Our brains detect that imbalance which is what triggers the breathing reflex. The more CO2 you give them- the higher the level of carbonic acid- and the stronger the reaction. In other words- it's not lack of oxygen that causes us to want to breathe- it's too much CO2.
If you fill a room with only Nitrogen, or just place a non-rebreather mask on their face and connected to a Nitrogen tank- they will simply fall unconscious and die within a couple of minutes. There is no pain, and no panic.
It is, in fact, a painless and quick death.
I've used FreeBSD for as long as I can remember. My desktop is _still_ FreeBSD.
Having said that- nothing performs under high load situations like Solaris. I was not happy with Solaris in the 7, 8 or 9 days (Mostly because they kept changing things- seemingly at random), but Solaris 10, and especially OpenSolaris, have been rock solid and cool as can be under high load.
My current OpenSolaris heads are running on Dell 2950's with 32 GB of RAM, Myricom Dual Port 10Gb ethernet nics, 8 lane (2 x4) 3Gb SAS controller, dual port LP10000exDC Emulex 2Gb/s Fiber Channel cards, and dual 3GHz quad core processors. They're connected to 6 Promise vTrak J610s Dual controller disk shelves. The shelves have 4 32GB write optimized SSD's in a mirrored/striped config spread across both controller channels for our ZIL. There are 80 15k SAS drives and 8 256GB Read optimized SSD's as the L2ARC.
Both heads are connected to the shelves in an active/passive configuration in which the second head can immediately pick up if the first one fails.
We use COMSTAR to export ZFS volumes via iSCSI (over 10Gb ethernet to our VMWare servers) and Fiber Channel (To our older Sun QA Farm).
You can scale ZFS pretty amazingly large- you just have to architect the system.
We've got 80 15k SAS spindles spread across two 4 lane 3Gb/s (12 Gb/s x 2) controllers. We've got 4 32GB write optimized SSD's in a mirror+stripe configuration, two disks on each channel for our ZIL. We've got 32 GB of RAM for dedupe and our L1-ARC. We've got 8 read optimized 256GB SSD's as our L2-ARC read cache. There are four hot spares in our system- 2 300GB and 2 600GB to match the two sizes of disks that we have.
The system absolutely flies. We've got dual 10Gb ethernet links in each of the two server heads. The heads are active/passive and both are connected to all the disks at all times via active/active SAS switches in the disk shelves (Promise vTrak J610S Dual's)
We use COMSTAR for iSCSI and FiberChannel target capabilities. (The fiber channel is 2Gb for our old Sun QA servers.
All of this still came out to less than 1/4 of what NetApp wanted for a similar system.
Sorry :)
That's _exactly_ what I do. I carry my keys on me- and my lock picks in my bag- if I lose one or the other- I'm still ok :)
Routing and ISP's are huge topics- what are you trying to do?
The main problem with routing isn't bandwidth- anyone can pump enough 1500 or 9000 byte frames per second to fill a gigabit pipe. The problem is when you have lots of small packets. At that point, dedicated routing hardware with a high-speed TCAM becomes really important.
What kind of line cards do you need? ADSL? Ethernet? OC12?
What kind of services do you need to run? BGP? OSPF?
What kind of bandwidth are you going to be pushing?
The basic Brother 4040CDN I just bought is a color laser with 4 individual toner cartridges, duplex, 21ppm and networking for $360- and it will last me a decade at least. It's not a photo printer, but if all you need is color it's a better choice than an overpriced inkjet.
Funny- but lots of houses in Levittown had copper radiant floor systems that failed and were attributed to the copper. We also have houses in Florida and Long Island developing pinhole leaks- either due to impurities in the copper, or environmental degradation. Either way- a single, jointless run of PEX is a hell of a lot safer bet (especially considering it can flex) than a copper pipe that can corrode, has lots of joints, and can't flex.
This is actually a law in a lot of states. The actual wording varies from "Left lane is for passing only" to "you can be in the left lane but must not impede faster traffic" to "you can be in the left lane, but must be driving at the speed limit and should not impede faster traffic."
Stop worrying about what's going to make you the most money and figure out what you enjoy. An MBA that hates his job is worthless. A computer scientist that isn't passionate about math and theory is worthless. An IT guy that isn't obsessed with all things tech will never be as good as the guy that is.
Figure out what you love doing and do that. If you really love it you'll be better at it. The best people in any field always make plenty of money.
As an aside- the last thing this world needs is more lawyers. The second to last thing this world needs is more MBA's.
Volkswagen has been making a diesel Golf that gets 60+ miles per gallon for several years now. The only "news" here is that the European division of Ford isn't as incompetent as their American division.
Any woman that is more concerned with electing a woman to the presidency, than they are with electing someone who represents their values, should lose their right to vote. It's idiotic in the extreme. If your grandmother wants to elect a fossil to the presidency because she's old, hey be that stupid. If she wants to elect McCain because he nominated a woman for the vice presidency, then your grandmother is an idiot who is playing right into McCains plans.
He nominated Palin because he wanted to tap the Hillary vote. If you think women are too stupid to see past such a ruse- then either you're also an idiot, or this country is hopelessly lost.
Did you even bother to read the original post? The person said their client wanted read-only access. The problem isn't the client messing up data- it's the client bringing the server to its knees with an abysmally inefficient query. Oracle has some features to limit the damage a user could do- but the only truly safe option is a read-only replica.
The person also wrote that the customer did not want 24 hour old data (you really didn't bother to read the post did you?) so your cube is a useless idea.
If you think it's a good idea to give clients direct access to your production database then please send me over your resume so I can make sure it goes on our "Never, ever hire this person" board.
They are incompetent- completely and utterly incompetent. They know only what encase or another piece of forensic software tells them. If the disk blocks have been rewritten a couple of times- they're not going to find it. They're not going to break AES unless you've done something stupid and left the key laying around.
The real bitch of it is- these guys never get challenged properly- especially in child porn cases. (Thank John Walsh- Adam's Law is absurd). They can claim whatever they want and the defense is basically helpless. The defense is not allowed to have their own copy of the drive to do forensic analysis on. They have to do it at the FBI lab with FBI equipment and with FBI goons hanging over their shoulders. If the FBI finds "overwritten" evidence- there is no good way to challenge that. It's your word against theirs.
Chain of custody? HAH! I've watched these guys leave crime scenes with drives under their arms, I've watched them run programs and click around a system they suspect of containing illegal material. No effort made to prevent trojans or other programs from covering their tracks. No effort made to preserve the state of the system. It's laughable.
And no- I wasn't a target. I did "forensic" analysis for years and got sick of watching these people make a mockery of my profession. (I put forensic in quotes because there is nothing scientific about these analyses- they are the best guesses of someone who may or may not be even remotely qualified to give an opinion).
What the hell do fingerprints have to do with this? You can't find a persons relatives via a finger print. You can't say "hey- think fingerprints looks almost like this other guys, so it much be a relatives!"
With DNA, you're using a DNA sample from a crime scene and matching it to a known criminals DNA to find a relative.
What you're suggesting is using a fingerprint from a crime scene, matching it to a known criminal, and then using that to find the persons relatives. That doesn't make the slightest bit of sense. If the fingerprint matches, you know your criminal. If it doesn't, you've got to keep investigating. Who they're related to isn't exactly important.
In this case we're talking about casting suspicion over people simply because their DNA is close to someone else's- that's frightening.
Ahh well, I had been led to believe they had written their own engine. -sirket
I love postgres- use it to handle millions of queries per day- but let's not kid ourselves: postgres replication in the form of slony blows. I have a master DB and it plublishes to a dozen read only databases. Managing that with slony just plain sucks. The simple fact is that setting up replication with mysql is dirt simple and that's part of why people use it.
I'm thinking of giving EnterpriseDB and their custom replication engine a try.
-sirket
Those are the questions we can't get people to answer now. There are plenty of harder questions we can ask, but we'd be happy if we could find people who can answer those.
If you put 5 years of netwroking experience we're going to ask you about the 5 types of OSPF LSA's and what a "Not So Stubby Area" is. If you admin a website we're going to ask you what the primary differences between HTTP 1.0 and 1.1 are. If you put security we're going to expect you to explain public key cryptography in some reasonable detail. If you put programming down you damned sure better be able to explain NP completeness and know how to write a compiler. There are tons of questions to ask, including plenty of trick questions, but like I said- I'd be happy to find someone who can answer the basics at this point.
-sirket
Does DNS use TCP or UDP? Explain. (The correct answer is not "TCP is used only for zone transfers.")
/etc/services, that's a good answer too.)
What port does service XYZ run on where XYZ is anything from SMTP or SSHD to NFS. (It's not about the right answers- if someone says I'd have to check
What information is stored in an inode?
Explain the boot process, from the BIOS on up to a fully running linux kernel.
What is an initrd and why does it exist?
You have a system with ABC problem- how would you troubleshoot it?
How would you stop a RedHat system (or any other) with two network interfaces on two different networks from routing traffic between those interfaces? Do this on a running system, and do it on startup. There are several ways. Name as many as you can.
What is a file's link count and what does it mean?
What does the link count mean with respect to a directory? How many links would a particular directory have? Why?
Explain routing and netmasks.
Etc.
-sirket
I work for a large Unix (primarily Linux) shop and we've been interviewing people for a year now. We're in NYC so there should be plenty of candidates- and there are. 1 in 10 passes our basic screening test (10 questions and not difficult). We then interview the ones we think are competent in person. It's a tough interview- but you don't need to get everything right- just handle yourself well and show some thought. Hell I'd settle for someone saying "I'd need to read the man page in response ot every question we ask).
Only two candidates have passed, and both of those only barely. We considered them entry level people and were willing to pay them entry level rates ($70k - $80k) but they wanted 6 figures. These are people that would have required significant training to reach a level we consider good.
Maybe we just have high standards but we haven't seen a single serious candidate in the last year. Say what you want but the market is definitely tight- at least for a serious Unix admin.
The one thing that has to stop is people lying about their experience. Do not list "networking experience including BGP" if all you've ever done is type "router bgp blah" and "network blah." If you don't know how OSPF works and can give me a decent explanation, then don't list OSPF. Don't list X years of RedHat experience and RedHat certification if you can't tell me how to find out what version of RedHat a box is running. Don't tell me Linux experience if you're not familiar with basic troubleshooting tools like iostat and vmstat. Don't list Linux if you can't tell me something as simple as how to turn off routing in your distro of choice. Or if you don't know- at least don't suggest stupid things like disabling one of the two interfaces or adding (or removing) a route. It's really annoying to hear such stupid answers.
-sirket
I second the Voicepulse Connect recommendation. Their web page sucks, a lot of information is missing, but in the end they're not doing that to hit you with hidden fees, their web department just looks to be incompetent :)
JHust to clarify- the company mentions possibly going to 28 stacked chips which would be 448 gigabits (not gigabytes) of storage- or about 56 GB of space. Now as flash chips grow in size- this could double (assuming 32 Gb NAND chips which are becoming available) to 96 or 112 GB of storage or more (assuming larger chips).