Well that's good thinking, I remember hearing about someone who bought connectivity from two different companies to have a redundant circuit, unfortunately both companies bought traffic through the same cable upstream, so every spring their connection got eaten by a backhoe.
Can we please stop fucking around with this myth? I have gone through 5 ATI boards (latest greatest ftw.) never had problems. You are more than likely running some old crap, and yes they sucked at drivers back then, but this is now and they just work [tm]*
Actually when you got a hot fail over why not use it for some load balancing?
And the solution is to use virtual servers, some of them support packing down a server and moving it to another physical server - that way you can power down half or more of your physical hardware but still keep all "servers" online. (Provided the systems aren't doing much at night)
"There is lots of cpu cycles left" Uhm. You are losing 30% cpu cycles, that is quite a lot. Yes there is amble power left for your office apps etc. but original poster says he is doing high performance computing - losing 30% of your throughput for reading data is a lot!
You probably need to get some counseling on your fetish for open source when you with absolutely no evidence of restricted drivers even being present on said system starts blaming them.
The law might actually prevent free wifi, here in Denmark for instance you are liable for whatever is going on on the connection, leaving it open intentionally can get you into a lot of trouble. Any cafe running wifi will make damned sure they know who was using the connection at what point (ig. require authentication and most likely creditcards).
Was in Prague and London this summer and both cities where nicely covered with wifi, however, subjectively the service was much much better in Prague than London - you have to look hard in Prague for a cafe without free wifi. Now the amount of AP's might be higher in London, but travelers/citizens ability to get online seemed quite a bit better in Prague.
Sorry but you have absolutely no idea what you are talking about.
Yahoo and Google do not have any mission critical data on those systems. Yes searching is important, but if a major clusterfuck happened they can just go get the data again. On top of that they are purely for reading data, MySQL replicates just fine for reading data, as long as you don't mind losing a couple of rows whenever a write master drops. This has been proven many times by big sites.
Regarding PgSQL; there are no native replication options, there are no native clustering options and they have many times stated they don't want to focus on it. This is just fine, their project, their decision. Other projects have tried making replication / clustering using PgSQL, but none work optimal, most work like MySQL where the slave is called a hot standby, but at best is warm - or using a middleware layer to control the cluster. Again this might work in some cases where you don't really care about your data - for instance search engines - if you lose your data you just go grab it again.
Highly intensive write applications with extremely critical data cannot rely on these parameters, doing so is downright insane. Talking about Petabyte sized databases must be doing the right thing is downright silly, if you have no idea about what they are doing with the data and how it arrives - and more importantly - what failure rate is acceptable you just can't compare it.
Are you saying that because top 10 largest databases (might be) are running PgSQL, I must be wrong because it's right for them, then it must be right for me?
As I've said time and time again, there are no FOSS databases that can compete with closed source software when it comes to fail over handling. PgSQL is an extremely nice full featured database, one thing it doesn't have - and most important to our organization - is hot standby option / clustered option. We don't care about petabyte sized databases, we care about speed and reliability.
The only time I would need to use Wikipedia is when I'm unsure of the subject and thus not able to know if what is said is true. Looking at the stuff I know about I can be pretty sure they can't be trusted; how could this ever be useful?
Quite a lot say; "well it is a good jumping point, get some pointers" - but again that makes no sense, if your sources are tainted to begin with you can't be sure anything can be trusted. Just look at many historian citing each other - often you end up with a nice circle where all are citing each other and most of it is BS.
God I love it when someone polishes his own ego by pointing out the obvious.
The whole point of the discussion was if open sourceness would make a good government. Wikipedia is open source and it is ruled by a government like structure, just not a democratic one. They are run like a good old dictatorship.
That's the whole point of the suit. Too bad the ruling wont have influence here in Denmark, it is very annoying that some company has copyright on whats on TV.
Slashdot Mirror
Well that's good thinking, I remember hearing about someone who bought connectivity from two different companies to have a redundant circuit, unfortunately both companies bought traffic through the same cable upstream, so every spring their connection got eaten by a backhoe.
It's called backhoes and they will eat up your uptime in one scoop.
That would require you to brute force the other players player key, which is of course pretty much impossible...
"ATI drivers don't even fucking work"
Can we please stop fucking around with this myth? I have gone through 5 ATI boards (latest greatest ftw.) never had problems. You are more than likely running some old crap, and yes they sucked at drivers back then, but this is now and they just work [tm]*
*under ubuntu.
Yeah me too, my wireless has no problems under Ubuntu. Just installed Ubuntu and I was good to go.
Actually when you got a hot fail over why not use it for some load balancing?
And the solution is to use virtual servers, some of them support packing down a server and moving it to another physical server - that way you can power down half or more of your physical hardware but still keep all "servers" online. (Provided the systems aren't doing much at night)
"There is lots of cpu cycles left"
Uhm. You are losing 30% cpu cycles, that is quite a lot. Yes there is amble power left for your office apps etc. but original poster says he is doing high performance computing - losing 30% of your throughput for reading data is a lot!
This election is pretty much sealing the fate of the western world as we know it; we are quite a lot of people very very interested in the results.
You have no idea what 5 9's are all about if you think that one box can handle it.
Yeah, blame it on closed source.
You probably need to get some counseling on your fetish for open source when you with absolutely no evidence of restricted drivers even being present on said system starts blaming them.
The law might actually prevent free wifi, here in Denmark for instance you are liable for whatever is going on on the connection, leaving it open intentionally can get you into a lot of trouble. Any cafe running wifi will make damned sure they know who was using the connection at what point (ig. require authentication and most likely creditcards).
Was thinking something fishy was going on.
Was in Prague and London this summer and both cities where nicely covered with wifi, however, subjectively the service was much much better in Prague than London - you have to look hard in Prague for a cafe without free wifi. Now the amount of AP's might be higher in London, but travelers/citizens ability to get online seemed quite a bit better in Prague.
I like the way the register puts them - Record industry Ass. of *. Perhaps we should use (_*_) to denote all countries?
Sorry but you have absolutely no idea what you are talking about.
Yahoo and Google do not have any mission critical data on those systems. Yes searching is important, but if a major clusterfuck happened they can just go get the data again. On top of that they are purely for reading data, MySQL replicates just fine for reading data, as long as you don't mind losing a couple of rows whenever a write master drops. This has been proven many times by big sites.
Regarding PgSQL; there are no native replication options, there are no native clustering options and they have many times stated they don't want to focus on it. This is just fine, their project, their decision. Other projects have tried making replication / clustering using PgSQL, but none work optimal, most work like MySQL where the slave is called a hot standby, but at best is warm - or using a middleware layer to control the cluster. Again this might work in some cases where you don't really care about your data - for instance search engines - if you lose your data you just go grab it again.
Highly intensive write applications with extremely critical data cannot rely on these parameters, doing so is downright insane. Talking about Petabyte sized databases must be doing the right thing is downright silly, if you have no idea about what they are doing with the data and how it arrives - and more importantly - what failure rate is acceptable you just can't compare it.
Your post makes little sense to me.
Are you saying that because top 10 largest databases (might be) are running PgSQL, I must be wrong because it's right for them, then it must be right for me?
As I've said time and time again, there are no FOSS databases that can compete with closed source software when it comes to fail over handling. PgSQL is an extremely nice full featured database, one thing it doesn't have - and most important to our organization - is hot standby option / clustered option. We don't care about petabyte sized databases, we care about speed and reliability.
Really? We pay for our database software, quite a lot of people would argue there are free alternatives, might not be just as good, but they are free!
Well you can keep your free alternatives until they are as good.
Also it is the price tag of the prototype ($50.000 for a prototype is cheap), they expect the resale price to be a lot lower.
Where do we sign up for your program?
Yeah, because getting told "the source is right there, fix it yourself" is so nice after you got telefragged...
Is it now?
The only time I would need to use Wikipedia is when I'm unsure of the subject and thus not able to know if what is said is true. Looking at the stuff I know about I can be pretty sure they can't be trusted; how could this ever be useful?
Quite a lot say; "well it is a good jumping point, get some pointers" - but again that makes no sense, if your sources are tainted to begin with you can't be sure anything can be trusted. Just look at many historian citing each other - often you end up with a nice circle where all are citing each other and most of it is BS.
God I love it when someone polishes his own ego by pointing out the obvious.
The whole point of the discussion was if open sourceness would make a good government. Wikipedia is open source and it is ruled by a government like structure, just not a democratic one. They are run like a good old dictatorship.
That's the whole point of the suit. Too bad the ruling wont have influence here in Denmark, it is very annoying that some company has copyright on whats on TV.
Uhm... so if there is a feature in anything you just fork it every time?
Just because you "can" doesn't mean it's feasible. My solution is to stay clear of Wikipedia...
Wikipedia has proven time and time again that "openness" will be corrupted just as easily as anything else.
Uhm...
So you want to spend $120 a year on crap vs. $100 for something good?
Now that makes a whole lot of sense..