I've used velcro (in rolls) for well over a decade everywhere from my office to the datacenter, and I've yet to find anything easier, cheaper or that works better than a plain old roll of velcro and a pair of scissors.
Gotta disagree with you. For a desk I think one of the best solutions is a roll of velcro and some screw in plant hooks on the back of a desk. Very easy to wrap cable and remove when you need to make changes, and a lot easier to deal with than cable ties. But definitely don't use zip ties, god what a nightmare.
Ask Google. They've only got 20,000 employees but estimates put the number of servers well into the 6 figure range.
It really depends on the business model, specifically if you're offering services on the Internet, you've probably got a small administrative staff managing a huge number of servers. Or what about companies with big render farms?
Here we have about 80 servers (~60 physical, the rest virtualized) and about 600-700 workstations. We also have over 6,000 employees.
Because the rest of the world doesn't agree with your views. You're a game developer, you're going to be probably more of a purist than the end-gamer (is that a term?). That's like saying, why do we have Blu-Ray? I just care about the plot and history of the movie.
The obvious question is, why shouldn't gamers expect it all? Compelling gameplay, great story lines and breathtaking graphics?
Those seem like at least realistic assumptions to me, at which point you have to wonder - which would you chose? The one with a decade of engineering from HP behind it or Zenoss? I've worked with Zenoss and I've seen OpenView and I made up my mind before I finished writing this post:)
If we assume 50% for Zenoss we need to assume the same for HP, who is very aggressive on their pricing in general ($work is a 100% HP shop for systems, desktop and datacenter). In which case, you can buy HP OpenView outright for the price of one year of Zenoss support.
Now, unless HP charges as much for support as they do for the software (obviously they don't) then the second year OpenView becomes cheaper than Zenoss. Which is just hilarious since HP OpenView is lauded constantly for being absurdly overpriced -- because it is, just like Zenoss's support prices.
Just for fun. Zenoss, highest level of support for 5k nodes is $900,000 per annum. Let's throw in a 25% discount which gives us $675K. So, in two years of Zenoss support you could buy the complete OpenView suite at full retail list prices. Let's assume the same discount rate for OpenView and we get right at an even $1M. So the ROI for purchasing OpenView (an obviously superior product) is only 17 months.
The very concept that HP OpenView is cheaper than the support for a open source project literally makes my stomach turn.
Are we comparing software prices to annual support costs? Those are ANNUAL costs for Zenoss. Even a 50% discount on 5,000 nodes would be a quarter of a million dollars for their LOWEST level of support.
Next, Zenoss's offering is not comparable to OpenView, sorry. HP OpenView is a massively more mature and robust product, and is really in a totally different league than Zenoss.
Zenoss's commercial support prices are hilarious, I mean, literally, hilarious. The CHEAPEST support (silver) is $100 per managed host (including virtualized hosts) most expensive (platinum) is $180 per node. So your 5,000 hosts would be $500,000-$900,000 per year in support.
Yes. Seriously.
The other problem I have with Zenoss is the reporting is basically non-existant. It may sound like I'm being hyper-critical, but it's only because I've looked at Zenoss and I so wanted it to be the NMS for me (I particularly like the fact that it's both open source and written in python) but at this point I just don't think it's going to work.
We use What's Up Gold from ipswitch right now, but we're only monitoring a few hundred hosts. It's slow, runs on windows, requires ms sql, but it's surprisingly full featured and gets the job done I suppose. Oh and its $900.
"If you lose your RAID array, you shouldn't lose any data. If you do, your backup strategy sucks."
You're nearly correct. You shouldn't lose any data since the last full/incremental/differential backup. Typically never more than 23 hours and 59 minutes:)
That's a pretty dated philosophy. These days backup software automatically checks tape integrity and tapes are overwritten pretty regularly in most situations.
RAID1 will *probably*, to an exceptionally high degree of likelihood, protect you from a single failed drive. I've seen dozens and dozens of drives fail in a RAID1 configuration, never once have I seen a RAID1 set fail entirely from a single or two simultaneous disk failures. I'm not implying it's impossible, just that it's very VERY unlikely, much more so than some people like to try and imply.
Well if you encrypt the files the integrity is basically guaranteed. If the company should suddenly get bought you just find a new company and start backing your files up there. You're also assuming that they'll just flip the switch and turn the service off overnight with absolutely no warning, and even then, unless before you can upload to a new backup service you lose some files, you'll be fine.
I think we have a language barrier here. Most home users don't NEED RAID. That's a fact. I can prove it because most home users don't have RAID setups. If home users NEEDED RAID then logic would dictate that most home users would HAVE RAID. Again, for the final time I will repeat my first quote.
Actually 70 double-layer DVDs, which would fit in a small CD wallet. I've got several CD/DVD cases on my shelf that are about 3x8x11 inches and hold about 250 DVDs. Not that I'm recommending it, just, as usual, you make it sound much more bleak than it is in actuality.
My setup consists of local backups to another set of hard drives along with a offsite storage (on a colocated server, encrypted). So if my house burns down or is hit by a meteorite and I lose both of my local copies of my data then I'll need to spend less than 5 days to download 1TB (@ 20Mb/s 4.85 days by my math). Of course personally I have nowhere near a terrabyte. And as you claimed, 600GB would only take about 3 days. I don't see what the problem is? If you need something sooner just download that first. I don't think anyone needs immediate access to the entire storage system, most is for archiving purposes.
I'm definitely open to better alternatives if you've got one.
Did you even read my post? I pointed out specifically that it was for availability and PERFORMANCE. If you need performance, great, buy a few more disks for a RAID setup. So don't start your post with "actually," like I said anything counter intuitive to that at all.
I've been using computers since before Windows existed and I don't remember how to use Win95 either, because like the rest of us I've permanently blocked those painful years from my memory.
"My experience is that NAS is a _lot_ slower than local storage for many applications. "
It's really about storage requirements. If my Internet connection is only 20mb/s and I just want to stream some movies at 50mb/s (1080p) then does it really matter if the network is "only" 100mb/s?
Slashdot Mirror
I've used velcro (in rolls) for well over a decade everywhere from my office to the datacenter, and I've yet to find anything easier, cheaper or that works better than a plain old roll of velcro and a pair of scissors.
"I have done this with 20+ devices in a 72u rack (over 6 feet of eq...)"
I think you mean a 42U rack . One Rack Unit (1RU) is 1.75 inches, which gives you a 42U rack at 73.5" or just over 6 feet.
http://en.wikipedia.org/wiki/Rack_unit
Gotta disagree with you. For a desk I think one of the best solutions is a roll of velcro and some screw in plant hooks on the back of a desk. Very easy to wrap cable and remove when you need to make changes, and a lot easier to deal with than cable ties. But definitely don't use zip ties, god what a nightmare.
We run Open-AudiT here as well, it's actually pretty awesome. Yes, it will run on Linux.
Ask Google. They've only got 20,000 employees but estimates put the number of servers well into the 6 figure range.
It really depends on the business model, specifically if you're offering services on the Internet, you've probably got a small administrative staff managing a huge number of servers. Or what about companies with big render farms?
Here we have about 80 servers (~60 physical, the rest virtualized) and about 600-700 workstations. We also have over 6,000 employees.
Because the rest of the world doesn't agree with your views. You're a game developer, you're going to be probably more of a purist than the end-gamer (is that a term?). That's like saying, why do we have Blu-Ray? I just care about the plot and history of the movie.
The obvious question is, why shouldn't gamers expect it all? Compelling gameplay, great story lines and breathtaking graphics?
Those seem like at least realistic assumptions to me, at which point you have to wonder - which would you chose? The one with a decade of engineering from HP behind it or Zenoss? I've worked with Zenoss and I've seen OpenView and I made up my mind before I finished writing this post :)
If we assume 50% for Zenoss we need to assume the same for HP, who is very aggressive on their pricing in general ($work is a 100% HP shop for systems, desktop and datacenter). In which case, you can buy HP OpenView outright for the price of one year of Zenoss support.
Now, unless HP charges as much for support as they do for the software (obviously they don't) then the second year OpenView becomes cheaper than Zenoss. Which is just hilarious since HP OpenView is lauded constantly for being absurdly overpriced -- because it is, just like Zenoss's support prices.
Just for fun. Zenoss, highest level of support for 5k nodes is $900,000 per annum. Let's throw in a 25% discount which gives us $675K. So, in two years of Zenoss support you could buy the complete OpenView suite at full retail list prices. Let's assume the same discount rate for OpenView and we get right at an even $1M. So the ROI for purchasing OpenView (an obviously superior product) is only 17 months.
The very concept that HP OpenView is cheaper than the support for a open source project literally makes my stomach turn.
Are we comparing software prices to annual support costs? Those are ANNUAL costs for Zenoss. Even a 50% discount on 5,000 nodes would be a quarter of a million dollars for their LOWEST level of support.
Next, Zenoss's offering is not comparable to OpenView, sorry. HP OpenView is a massively more mature and robust product, and is really in a totally different league than Zenoss.
Zenoss's commercial support prices are hilarious, I mean, literally, hilarious. The CHEAPEST support (silver) is $100 per managed host (including virtualized hosts) most expensive (platinum) is $180 per node. So your 5,000 hosts would be $500,000-$900,000 per year in support.
Yes. Seriously.
The other problem I have with Zenoss is the reporting is basically non-existant. It may sound like I'm being hyper-critical, but it's only because I've looked at Zenoss and I so wanted it to be the NMS for me (I particularly like the fact that it's both open source and written in python) but at this point I just don't think it's going to work.
We use What's Up Gold from ipswitch right now, but we're only monitoring a few hundred hosts. It's slow, runs on windows, requires ms sql, but it's surprisingly full featured and gets the job done I suppose. Oh and its $900.
"If you lose your RAID array, you shouldn't lose any data. If you do, your backup strategy sucks."
:)
You're nearly correct. You shouldn't lose any data since the last full/incremental/differential backup. Typically never more than 23 hours and 59 minutes
And when you accidentally delete that file you've been working on for a week you'll be glad you had your backups :)
That's a pretty dated philosophy. These days backup software automatically checks tape integrity and tapes are overwritten pretty regularly in most situations.
Linux software raid is fantastic.
You're forgetting the most common case - "oh shit did I just delete the wrong directory?"
RAID1 will *probably*, to an exceptionally high degree of likelihood, protect you from a single failed drive. I've seen dozens and dozens of drives fail in a RAID1 configuration, never once have I seen a RAID1 set fail entirely from a single or two simultaneous disk failures. I'm not implying it's impossible, just that it's very VERY unlikely, much more so than some people like to try and imply.
For anyone looking at drobo I highly recommend checking out unRAID: http://www.lime-technology.com/
Well if you encrypt the files the integrity is basically guaranteed. If the company should suddenly get bought you just find a new company and start backing your files up there. You're also assuming that they'll just flip the switch and turn the service off overnight with absolutely no warning, and even then, unless before you can upload to a new backup service you lose some files, you'll be fine.
I think we have a language barrier here. Most home users don't NEED RAID. That's a fact. I can prove it because most home users don't have RAID setups. If home users NEEDED RAID then logic would dictate that most home users would HAVE RAID. Again, for the final time I will repeat my first quote.
MOST home users don't NEED RAID.
Get it?
Actually 70 double-layer DVDs, which would fit in a small CD wallet. I've got several CD/DVD cases on my shelf that are about 3x8x11 inches and hold about 250 DVDs. Not that I'm recommending it, just, as usual, you make it sound much more bleak than it is in actuality.
My setup consists of local backups to another set of hard drives along with a offsite storage (on a colocated server, encrypted). So if my house burns down or is hit by a meteorite and I lose both of my local copies of my data then I'll need to spend less than 5 days to download 1TB (@ 20Mb/s 4.85 days by my math). Of course personally I have nowhere near a terrabyte. And as you claimed, 600GB would only take about 3 days. I don't see what the problem is? If you need something sooner just download that first. I don't think anyone needs immediate access to the entire storage system, most is for archiving purposes.
I'm definitely open to better alternatives if you've got one.
Did you even read my post? I pointed out specifically that it was for availability and PERFORMANCE. If you need performance, great, buy a few more disks for a RAID setup. So don't start your post with "actually," like I said anything counter intuitive to that at all.
I've been using computers since before Windows existed and I don't remember how to use Win95 either, because like the rest of us I've permanently blocked those painful years from my memory.
"My experience is that NAS is a _lot_ slower than local storage for many applications. "
It's really about storage requirements. If my Internet connection is only 20mb/s and I just want to stream some movies at 50mb/s (1080p) then does it really matter if the network is "only" 100mb/s?