I'll probably get flamed for this but I have to say it anyway...
Why does anyone still expect evolution in our society? With the social system and the way our economy works there is no reason for evolution anymore. If you take a pack of lions... The top is the strongest animal, then the second tier is the ones that are almost as strong and so on. Now I look at where I work - the richest and most powerful guy has his job cause he started almost at the top and had the right backing... The next level down are all his friends - most of them completly incompetent idiots. Evolution? No thanks!
Now the other side - and that's the really scary one - since when do we weed out bad genes? Today most people die a natural death, no matter if they were stupid, disabled or had any other issues. In the past, those would have been the first to get killed by lack of food, deciese or wild animals. That kept the gene pool cleaner. Today, they have kids just like everyone else - and that has severe negative impact on the human race.
I'm not saying that there is any ethical way of changing that or that it even should be changed, but if the topic of evolution comes up, most people just silently ignore these two facts most of the time...
Guys, just wait until Itanium is ready... This is just a 64bit extention to a 32bit extention to a 16 bit architecture... Itanium is where its really at! If you have any kind of high end workload you need Itanium! x64 can't handle it! I promise! The sheep are lying!!!!!!!!!!
Your writing sounds good - but if anyone puts "quality hardware" and "sun" in one sentence they show that they have never working with any of it... RSC resets for no apperant seasons on v480s, v20zs that don't open, split expander issues on F15k, memory issues on a v490 that took 4 service calls (on premium support level), v440s that overheat at an amazing 42 degrees celsius... that's just a few examples of the issues that we have had on the last generation of sun gear...
Hmm - I'm starting to wonder why people get excited about a box that has been on the market for years (its a newisys 4300 for gods sake, nothing new) and is going to be phased out shortly for a sun internal design? (well, not really sun - they bought the whole company this time than rather just oem the system)...
besides that, they run stable, fast and are all you'd expect from a 4way opteron... Just the NSV (network share volume) for the sp is pretty unusual... Also doesn't have a virtual CD, making it somewhat difficult if you can't network boot...
The only IT projects that failed that I know are the ones that have bad managers... Most of the time that means someone who doesn't listen to the people that do the actual work but there are other reasons...
So no one will ever have a soundcard in their computer? Or a CD player?
Not sure what qualifications they got but whoever puts D/A converters on a endangered list has proven that they don't have much understanding of electronics...
You got a few pointers here but no one that commented seems to have run a larger environment - cause there you have very very different issues. Some of them sound trivial but in a 600+ people shop its the small things that kill you.
First of, don't do symlinks even though they are possible... They always end up in a mess. Then make sure you provide them with a good way of changing their passwords (no matter what access method you're using) and see if you can't arrange it to force changes - otherwise people will share ids and its a pain. If nessecary, create functional IDs - that way you at least know its not one person but a group. Keep cvs on a Tru64(advfs) or Linux box... HP-UX or Solaris is a lot slower... solaris10 with zfs might be close but haven't tried that yet. Look for a few frontends (like cvsweb and so on) - they will help a lot for users that are new. Send out an email with the cvsbook link (link is on cvshome.org) and make sure you remind people that's a good reference. Offer them wincvs - windows users will not only have trouble adapting to command line but will screw up bigtime and after all its your job to fix it...
I've been on both sides of the table in too many interviews and honestly - the skills part is the most useless part of your interview. I have all but given up on that ever working. Specific skills are way too easy to study for - you never know what you'll get no matter how well someone passes the test. In my experience it has much more value if you try to get your answer indirectly... Like I was looking for a new sysadmin for my team. He has to answer questions of a ton of users that just walk up and at the same time get work done... So while doing something that takes long, he has to be able to either get interrupted and then go back to his work where he left off or he has to go tell people to hold on. Someone who gets interrupted and then doesn't know what he was doing will get lost. So while my college asked some questions that took a long time to explain I constantly interrupted (asking for more detail or other related short questions) and then watched the reaction of the candidate. One of the guys I interviewed got thrown off by it and wasn't able to ever go back to the question asked by my college - he would have failed in our environment...
Guess the point I'm trying to make is that you anyway never know what someone looks for - so don't bother studying or worrying what might happen. Relax, do your best and just try be yourself.
Unfortunately you omitted to say what size of installation you're talking about... I work at a large company and layout and maintain a development shop for like 600 developers... I prefer direct attached disks for a simple reason - if something fails, I have less people complaining about it:-)
Seriously, its not so much a speed issue but an issue of how you manage your environment. You can get enough performance out of all three solutions and there are other, more important things to look at.
Price: direct attached storage being the cheapest usually, then iscsi being a little more and fibre being by far the most expensive.
Redunancy: the higher the price, the more redundancy you usually get - its not a technology issue but a cost issue again... dasd is usually at the same level as iscsi since most manufacturers just give you a preconfigured computer with dasd and then sell it as iscsi... Fibre gives you the highest level of redundancy - because if you already go spend 5 million on symetrix frames its easy to justify the additional $400K for one more smaller connectrix. If you only talk about $4000 disk trays its hard to justify having an electrician comming in to give you redunant power for another $3000...
Clustering is another reason to go with fibre or iscsi - direct attached storage is usually a bad idea there...
Backups are another consideration.. Backup through ethernet is much slower than fibre tape units through a san...
Fibre and iscsi are not dedicated. So if I have a server with direct attached disks that happily does 2000 io/sec today, it will do so tomorrow unless that server has issues. If I have fibre or iscsi and lets say I run year end reports on another box, that heavy load there can drop my io rate to 20% or even lower.
Fibre is a bitch to configure the first time. direct attached disks are easy and iscsi is usually managable - but getting fibre right the first time is much more difficult. You'll break a fibre cable here cause its so delicate, have a bag gbic there and then can't get the san wwn masks right - and you just had the most frustrating 3 weeks of your life...
As you see, I would recommend worrying less about the performance and more about what you really need and what your environment looks like. If you just want performance for cheap, then local disks are unbeatable. Instead of spending money on expensive fibre or iscsi offload controllers buy tons of cheaper scsi cards, instead of running fibre and buying a fibre switch buy tons of disk drives. Most people make the mistake to worry about capacity - in reality, its the number of spindles you need to look at and the capacity or your storage is a result of that. Figure, each disk attached gives you about 100 i/o ops per second. If you need to do 5000 ops, you need 50 disks - no matter how they are connected. Figure a disk can get you 20 MB/sec - if you need 200MB/sec throughput, thats 10 disks. In that example, I need 50 disks then to satisfy my requirements. Next, take the max throughput and divide it by 3 - that's 66MB/sec for fibre then, 100MB for scsi and 33MB for iscsi over gigabit. So to run my 200MB/sec I'd need only 2 scsi channels, 3 fibre or 6 iscsi connections. Next of course can't have 50 drives on 2 channels - more than 3 disks drop the transfer rate dramatically... Since fibre and iscsi mask the physical spindles they don't care but I need to have 16 scsi controllers to really run the disks.
I work at a fortune 5 and my team supports things like E10K, F15K, Regatta, Superdome... The most important part is if you can trust the vendor. If you buy a $5mil frame, you want to make sure that not only will it be supported and the company will be around - but you also have to believe that the vendor will in the future be able to provide adequate patches and updates.
Nothing is more annoying than if you buy a big frame and then you find out that a silly little piece of software is no longer maintained. Or like HP announced today, that they are once again changing their HP-UX roadmap and once again proved that they can't be taken seriously if they predict anything further out than 3 months.
It all comes down to the simple fact that in the end, almost all of the big boxes are the same to the application. Sure, some have hard and some softpartitioning. Sure, you have different cpus, memory latencies and whatever - to the app it is just a bunch of system calls. But in the end, if you can't run your app on it, its useless, no matter how fast, redundant or whatever it is. We have completely moved away from selecting the box by its hardware properties. They are all sufficiently redundant and whatever. We go purely by how well the software we need to run is supported on the OS and if they have a roadmap that can be trusted.
I have such a card and it works just fine for me... the only downside is you don't have a link stability/power display (that would be from the second serial port).. other than that, works fine, very stable...
Off topic but muahahah - look at the history of the NeXTstep article on Wikipedia - someone messed it up and replaced NeXTstep with Hitler and stuff like that - now its locked:-)
If you got the money, go hardware... I have two systems that are identical with the exception that in one I broke one of the ide connectors of the 3ware controller and had to go software raid instead. I have 6 120GB disks in each, on separate channels... There is no problem with the stability and both setups survived at least 2 disk failures each.
Unfortunately the performance is another issue - I get three times the throughput on the 3ware controller. So the $300 I spent on the controller (used) was well worth it...
Even though most people will disregard that - check what moon phase you sleep the worst... For me, its a very strict pattern - can't sleep 3 days before and 2 days after the full moon... Other cultures have known this for thousands of years - maybe its time to rediscover things... I've changed my sleeping habits to going to bed as soon as I come home during the full moon period and sleep until the moon wakes me up. Enjoy google about this subject http://www.google.com/search?hl=en&ie=UTF-8&q=full +moon+accident+statistics&btnG=Google+Search
Why don't you just put up a small server? If you go outside your own network it will get so much slower and less comfortable to work with... I've been there and dropped it after a while - between downtime of your PC, downtime of the network and your storage provider its not really all that useful...
I ended up getting a storage router that allows me access from the outside if needed USR Storage Router 8200 its so much easier...
We already went though fax spam, email spam, telemarketers and of course everyone's favorite - junk in your snail mail box. I think its time that we come up with a more global view of things. A single list similar to the do-not-call list but that will allow you to get blacklisted for every kind of communications. I know many people have reservations like that spammers will use these lists as a source of valid email addresses, but you can get around that by allowing the user to select which one of their contacts they want on there...
Peter.
3ware + ATA hotswap trays
on
SATA vs ATA?
·
· Score: 2, Informative
We're running several servers with 3ware controllers and SATA drives where I work and while the controllers are great the SATA connectors suck. They are just too fragile. Everyone in my team who touched the setup - no matter how careful they were - ended up breaking a connector. If you have only one or two cables its alright - but once you end up having 8 or more and try to route them nicely you'll be in trouble.
If you're going for more than just 2 or 3 drives and want to go SATA you should go with one of the newer multilane connectors. One connector carries 4 SATA channels and for an array with 12 drives you only have to worry about 3 cables. That makes the cable layout much neater and the connectors are fairly solid.
Guess we're missing one angle - over here, most major IT shops in large corporations are (like the coporations themselves) grossly mismanaged. The company I work for requires us to buy EMC disks for development servers at roughly 45 times the cost of regular disks - even if you don't need the redundancy... So one of the development servers cost us 4500... the disk space for it was $900.000... Seriously, what does it matter if I spend 100K or just 25K on the three users of that server if the server itsself is the issue? Or to buy a new desktop - cost about 750 dollar - we spend $2500 figuring out if its really needed, if everyone approves it and so on... Add in that cost - and you have 95% of the reason why indian IT shops are cheaper than doing the same work in the US...
Little info and terms like "Photon Vacuum" make this thing sound like the next high end graphics card from the bitboys... If they had a usable product they would give you at least some information - especially since the design is according to them patent protected...
>NFS is stateless from the server's perspective. This is done so that the server doesn't have to track the state of a whole fleet of clients (and so that the server can pick up where it left off when it crashes and restarts).
Wrong - the nfsd is statless.. mounts, locks and stuff like that persists...
showmount -a is your friend... that will show you all clients that currently have a filesystem mounted... just move your shutdown, reboot or whatever out of the way, replace them by a wrapper script that checks if showmount -a returns any clients, and only executes the real shutdown when noone has the filesystems mounted... if they do you can always print out a list of workstations that have mounted filesystems...
I've solved the same issue for me with cheap plastic drawers from walmart and then lining them with the thin antistatic foam you find in like mainboard boxes and other places... Never had a card go bad on me in that storage so I guess it works good enough:)
Slashdot Mirror
I'll probably get flamed for this but I have to say it anyway...
Why does anyone still expect evolution in our society? With the social system and the way our economy works there is no reason for evolution anymore. If you take a pack of lions... The top is the strongest animal, then the second tier is the ones that are almost as strong and so on. Now I look at where I work - the richest and most powerful guy has his job cause he started almost at the top and had the right backing... The next level down are all his friends - most of them completly incompetent idiots. Evolution? No thanks!
Now the other side - and that's the really scary one - since when do we weed out bad genes? Today most people die a natural death, no matter if they were stupid, disabled or had any other issues. In the past, those would have been the first to get killed by lack of food, deciese or wild animals. That kept the gene pool cleaner. Today, they have kids just like everyone else - and that has severe negative impact on the human race.
I'm not saying that there is any ethical way of changing that or that it even should be changed, but if the topic of evolution comes up, most people just silently ignore these two facts most of the time...
Peter.
Guys, just wait until Itanium is ready... This is just a 64bit extention to a 32bit extention to a 16 bit architecture... Itanium is where its really at! If you have any kind of high end workload you need Itanium! x64 can't handle it! I promise! The sheep are lying!!!!!!!!!!
it seems to be a pre-production board or maybe even fake images. The real thing also has 2 memory slots...
. js p?motherboardId=321
http://www.viaembedded.com/product/epia_dp_spec
Go to VIA directly and you shall see...
Peter.
> - Excellent and consistent hardware quality
Your writing sounds good - but if anyone puts "quality hardware" and "sun" in one sentence they show that they have never working with any of it...
RSC resets for no apperant seasons on v480s, v20zs that don't open, split expander issues on F15k, memory issues on a v490 that took 4 service calls (on premium support level), v440s that overheat at an amazing 42 degrees celsius... that's just a few examples of the issues that we have had on the last generation of sun gear...
Peter.
Hmm - I'm starting to wonder why people get excited about a box that has been on the market for years (its a newisys 4300 for gods sake, nothing new) and is going to be phased out shortly for a sun internal design? (well, not really sun - they bought the whole company this time than rather just oem the system)...
besides that, they run stable, fast and are all you'd expect from a 4way opteron... Just the NSV (network share volume) for the sp is pretty unusual... Also doesn't have a virtual CD, making it somewhat difficult if you can't network boot...
The only IT projects that failed that I know are the ones that have bad managers... Most of the time that means someone who doesn't listen to the people that do the actual work but there are other reasons...
Peter.
So no one will ever have a soundcard in their computer? Or a CD player?
Not sure what qualifications they got but whoever puts D/A converters on a endangered list has proven that they don't have much understanding of electronics...
Peter.
You got a few pointers here but no one that commented seems to have run a larger environment - cause there you have very very different issues. Some of them sound trivial but in a 600+ people shop its the small things that kill you.
First of, don't do symlinks even though they are possible... They always end up in a mess.
Then make sure you provide them with a good way of changing their passwords (no matter what access method you're using) and see if you can't arrange it to force changes - otherwise people will share ids and its a pain.
If nessecary, create functional IDs - that way you at least know its not one person but a group.
Keep cvs on a Tru64(advfs) or Linux box... HP-UX or Solaris is a lot slower... solaris10 with zfs might be close but haven't tried that yet.
Look for a few frontends (like cvsweb and so on) - they will help a lot for users that are new.
Send out an email with the cvsbook link (link is on cvshome.org) and make sure you remind people that's a good reference.
Offer them wincvs - windows users will not only have trouble adapting to command line but will screw up bigtime and after all its your job to fix it...
Peter.
I've been on both sides of the table in too many interviews and honestly - the skills part is the most useless part of your interview. I have all but given up on that ever working.
Specific skills are way too easy to study for - you never know what you'll get no matter how well someone passes the test.
In my experience it has much more value if you try to get your answer indirectly... Like I was looking for a new sysadmin for my team. He has to answer questions of a ton of users that just walk up and at the same time get work done... So while doing something that takes long, he has to be able to either get interrupted and then go back to his work where he left off or he has to go tell people to hold on. Someone who gets interrupted and then doesn't know what he was doing will get lost. So while my college asked some questions that took a long time to explain I constantly interrupted (asking for more detail or other related short questions) and then watched the reaction of the candidate. One of the guys I interviewed got thrown off by it and wasn't able to ever go back to the question asked by my college - he would have failed in our environment...
Guess the point I'm trying to make is that you anyway never know what someone looks for - so don't bother studying or worrying what might happen. Relax, do your best and just try be yourself.
Peter.
Seriously, its not so much a speed issue but an issue of how you manage your environment. You can get enough performance out of all three solutions and there are other, more important things to look at.
As you see, I would recommend worrying less about the performance and more about what you really need and what your environment looks like.
If you just want performance for cheap, then local disks are unbeatable. Instead of spending money on expensive fibre or iscsi offload controllers buy tons of cheaper scsi cards, instead of running fibre and buying a fibre switch buy tons of disk drives.
Most people make the mistake to worry about capacity - in reality, its the number of spindles you need to look at and the capacity or your storage is a result of that. Figure, each disk attached gives you about 100 i/o ops per second. If you need to do 5000 ops, you need 50 disks - no matter how they are connected. Figure a disk can get you 20 MB/sec - if you need 200MB/sec throughput, thats 10 disks.
In that example, I need 50 disks then to satisfy my requirements. Next, take the max throughput and divide it by 3 - that's 66MB/sec for fibre then, 100MB for scsi and 33MB for iscsi over gigabit. So to run my 200MB/sec I'd need only 2 scsi channels, 3 fibre or 6 iscsi connections.
Next of course can't have 50 drives on 2 channels - more than 3 disks drop the transfer rate dramatically... Since fibre and iscsi mask the physical spindles they don't care but I need to have 16 scsi controllers to really run the disks.
Peter.
find, grep and vi is all you need! :-)
Nothing is more annoying than if you buy a big frame and then you find out that a silly little piece of software is no longer maintained. Or like HP announced today, that they are once again changing their HP-UX roadmap and once again proved that they can't be taken seriously if they predict anything further out than 3 months.
It all comes down to the simple fact that in the end, almost all of the big boxes are the same to the application. Sure, some have hard and some softpartitioning. Sure, you have different cpus, memory latencies and whatever - to the app it is just a bunch of system calls. But in the end, if you can't run your app on it, its useless, no matter how fast, redundant or whatever it is. We have completely moved away from selecting the box by its hardware properties. They are all sufficiently redundant and whatever. We go purely by how well the software we need to run is supported on the OS and if they have a roadmap that can be trusted.
Peter.
http://www.ka9q.net/5220.html
/power display (that would be from the second serial port).. other than that, works fine, very stable...
I have such a card and it works just fine for me... the only downside is you don't have a link stability
Off topic but muahahah - look at the history of the NeXTstep article on Wikipedia - someone messed it up and replaced NeXTstep with Hitler and stuff like that - now its locked :-)
Peter.
If you got the money, go hardware... I have two systems that are identical with the exception that in one I broke one of the ide connectors of the 3ware controller and had to go software raid instead. I have 6 120GB disks in each, on separate channels... There is no problem with the stability and both setups survived at least 2 disk failures each.
Unfortunately the performance is another issue - I get three times the throughput on the 3ware controller. So the $300 I spent on the controller (used) was well worth it...
Peter.
Maybe I'm dumb but I always thought apollo was bought by HP not sun ;)
Peter.
Even though most people will disregard that - check what moon phase you sleep the worst... For me, its a very strict pattern - can't sleep 3 days before and 2 days after the full moon...l +moon+accident+statistics&btnG=Google+Search
Other cultures have known this for thousands of years - maybe its time to rediscover things... I've changed my sleeping habits to going to bed as soon as I come home during the full moon period and sleep until the moon wakes me up.
Enjoy google about this subject http://www.google.com/search?hl=en&ie=UTF-8&q=ful
Peter.
--
http://www.dealrover.com
Why don't you just put up a small server? If you go outside your own network it will get so much slower and less comfortable to work with... I've been there and dropped it after a while - between downtime of your PC, downtime of the network and your storage provider its not really all that useful...
I ended up getting a storage router that allows me access from the outside if needed USR Storage Router 8200 its so much easier...
Peter.
We already went though fax spam, email spam, telemarketers and of course everyone's favorite - junk in your snail mail box.
I think its time that we come up with a more global view of things. A single list similar to the do-not-call list but that will allow you to get blacklisted for every kind of communications. I know many people have reservations like that spammers will use these lists as a source of valid email addresses, but you can get around that by allowing the user to select which one of their contacts they want on there...
Peter.
We're running several servers with 3ware controllers and SATA drives where I work and while the controllers are great the SATA connectors suck. They are just too fragile. Everyone in my team who touched the setup - no matter how careful they were - ended up breaking a connector. If you have only one or two cables its alright - but once you end up having 8 or more and try to route them nicely you'll be in trouble.
If you're going for more than just 2 or 3 drives and want to go SATA you should go with one of the newer multilane connectors. One connector carries 4 SATA channels and for an array with 12 drives you only have to worry about 3 cables. That makes the cable layout much neater and the connectors are fairly solid.
Guess we're missing one angle - over here, most major IT shops in large corporations are (like the coporations themselves) grossly mismanaged. The company I work for requires us to buy EMC disks for development servers at roughly 45 times the cost of regular disks - even if you don't need the redundancy... So one of the development servers cost us 4500... the disk space for it was $900.000... Seriously, what does it matter if I spend 100K or just 25K on the three users of that server if the server itsself is the issue?
Or to buy a new desktop - cost about 750 dollar - we spend $2500 figuring out if its really needed, if everyone approves it and so on... Add in that cost - and you have 95% of the reason why indian IT shops are cheaper than doing the same work in the US...
Peter.
Little info and terms like "Photon Vacuum" make this thing sound like the next high end graphics card from the bitboys... If they had a usable product they would give you at least some information - especially since the design is according to them patent protected...
>NFS is stateless from the server's perspective. This is done so that the server doesn't have to track the state of a whole fleet of clients (and so that the server can pick up where it left off when it crashes and restarts).
Wrong - the nfsd is statless.. mounts, locks and stuff like that persists...
showmount -a is your friend ... that will show you all clients that currently have a filesystem mounted... just move your shutdown, reboot or whatever out of the way, replace them by a wrapper script that checks if showmount -a returns any clients, and only executes the real shutdown when noone has the filesystems mounted... if they do you can always print out a list of workstations that have mounted filesystems...
I've solved the same issue for me with cheap plastic drawers from walmart and then lining them with the thin antistatic foam you find in like mainboard boxes and other places... Never had a card go bad on me in that storage so I guess it works good enough :)