If tables are stored in the InnoDB storage engine, mysqldump provides a way of making an online backup of these (see command below). This backup just needs to acquire a global read lock on all tables (using FLUSH TABLES WITH READ LOCK) at the beginning of the dump. As soon as this lock has been acquired, the binary log coordinates are read and lock is released. So if and only if one long updating statement is running when the FLUSH... is issued, the MySQL server may get stalled until that long statement finishes, and then the dump becomes lock-free. So if the MySQL server receives only short (in the sense of "short execution time") updating statements, even if there are plenty of them, the initial lock period should not be noticeable.
We use this and have never had any problems, but I guess we don't have these long running updates (at least not at 12:01 when our backup runs).
Sure, it's a script. Who cares? We used to backup by locking the database and copying the tables, but when we moved to InnoDB that didn't work any more. I timed it, and the mysqldump is just as fast to save and restore as the binary files when you use the "--tab" option.
The question to ask a_greer2005 is, "Would you have voted against Bush if you knew about this?" My guess is the answer is "no", because this would have meant voting for Kerry.
OTOH, I am quite sure there are parents who would see the FBI's action as a good thing - even Democratic parents.
This is classic wedge politics. Basically, anything that has more of a chance to cause voters on the other side to defect than your own voters qualifies.
The rare events make for more drama. The news is about drama, not NEWS.
If something is not a rare event, in what way is it news? I mean, what are you expecting?
ECONOMIC NEWS: Businesses Exchanging Goods and Services for Money FASHION NEWS: Fashion Elite Wear Warmer Clothes in Winter than Summer HEALTH NEWS: Ageing Linked to Death SCIENCE NEWS: Earth Revolving Around Axis Once Per Day
News tends to be about things that don't happen every day. Except news about Paris Hilton's latest scandal, but some things never get old.:)
Starting a wterm takes about a fifth as long as it does to fire up Konsole, for instance, and switching desktops is also much faster.
I am running on a 1.7 GHz laptop right now, and it takes less than one second to start up Konsole:
$ time konsole -e "/bin/echo"
real 0m0.603s user 0m0.409s sys 0m0.033s
Did you actually time the startup of the shell? Even if you did, do you really notice the 0.5 seconds of difference? For myself, KDE's "bloat" actually increases my efficiency, because I can create a new tab in Konsole rather than opening a new shell.
As for the desktop switching... the desktops switch faster than I can click - I tried to use the hotkeys to outrun the system (Ctrl-F1, Ctrl-F2,...), but it is faster than I am.
You are vastly overestimating the population of the Netherlands.:)
On a more serious note, the US started requiring a social security number (SSN, a taxpayer ID for the American pension scheme) for children claimed as "dependents" at a younger and younger age through the 1980's. This is part of the bizarrely complicated tax system in the US, but basically the government reduces your taxes for each child you take care of. The reason for requiring the SSN was to reduce people cheating by claiming they were supporting children who did not really exist. Supposedly it worked, but I can't be bothered to dig up statistics.
You're mistaken. When there is a threat, we mutate.
I am not a biologist, but I am pretty sure that while mutation does occur, it is rare and very few mutant genes survive to reproduce themselves even once. In general it is gene selection that causes most evolution rather than gene mutation.
By this I mean the case where a rare, but non-harmful, gene becomes useful due to a change in environment. In such a case the gene gains advantage over its alleles and becomes common.
At firms I worked with (telephony companies, usually), scheduled downtime is not included in downtime numbers. Of course, it depends on the SLA, but this is how it worked in the "5 nines" days of Ma Bell. 5 nines (99.999% uptime) was basically a myth.:)
Interesting. Experience tells us that eventually someone will develop the strategy of cheating against the other team mode players, thereby gaining even more of an advantage.
Ultimately the "team mode" question should be an unavoidable part of on-line playing. Do I collude with other players? If I do, how honest should I be?
We switched to InnoDB many years ago, because, frankly, MyISAM had locking problems.:)
We switched to 4.0 a couple of years ago, because we wanted the mysqldump that supported transactions and thereby allowed us to make a consistent backup without trickery.
BTW, InnoDB benched at about half the speed of MyISAM for us for a single query. However, we normally have concurrency, and the speed was only about 10% less in this case. And the worst-case performance was seconds rather than minutes, so it was a clear win.
We haven't seen any problems since we switched away from MyISAM (and had not with MyISAM for a long time either).
Indeed. I did some benchmarking of MySQL verus Microsoft SQL Server Express 2005 recently.
MSSQL works about 4x slower if you do not prepare queries properly. Even so, MySQL is significantly faster for the queries that matter in our application (but MSSQL is faster for certain other queries).
MySQL 4.1 and newer support prepared statements, if you really want them. (I'll benchmark that tonight, now that I think of it...)
Our update server gets about 14 updates per second. We've never had MySQL corruption on the box.
Both of our query servers get between 200 and 300 queries per second, in addition to mirroring the update server, and have never had MySQL corruption.
Our data is small (about 12 Gbyte), so maybe we're just lucky, but I think heavy activity is probably not the problem.
We *did* get some corruption once, about 4 years ago. At the time we were paying for MySQL support, pushed our files to the developers, and MySQL was fixed.
Are you sure it's not a bug in the application?
Re:Did that say signed vs. unsigned integer bugs?
on
OpenSSH 4.2 released
·
· Score: 1
Unfortunately due to bad language design, finding signed/unsigned problems is often a subtle problem in C.
Personally, I think the OpenBSD folks are doing things the hard way, by using an insecure language as the foundation of their work. That's the problem with C - you have to remember everything you learned over years of programming, all the time, or you risk making a mistake that can not only cause crashes, but ultimately compromise your entire machine, if not your entire network.
But having said that, I do use OpenSSH and occasionally OpenNTPd (on machines with interfaces that go up and down a lot).:)
In the eyes of most environmentalists, (or at least the environmental lobbying groups) nuclear power is more of a threat to the environment than global warming.
Possibly, but there simply is no alternative. It might be possible to set up vast areas of desert or ocean as solar collection points, but that would not be 100% environmentally neutral.
An hour with the Internet and a calculator will convince you that even an extremely efficient biomass program cannot replace current fossil fuel consumptions. Renewable resources such as wind and hydroelectric power can help, as can reduced consumption, but in the end we need nuclear power.
Rather, we need nuclear power if we want to continue the standard of living of the small fraction of the world's people that live in the west. The world needs it even more if it wants all of humanity to live in safety and comfort. (Unfortunately this is not really a primary goal for religious people, or for self-interested or nationalistic people, but it should be!):)
Some estimates are that fission would supply us with power for thousands of years. The main issues with fission are that you can build really scary weapons with the technology, and it produces waste that lasts effectively forever. These can be addressed, but I really believe that fusion power is the ultimate answer.
I've heard about how demoralizing it can be for our troops to be forced to kill children
The Nazi's developed gas as a method of killing Jews because of psychological problems appearing in the soldiers who were ordered to kill them.
I think the lesson learnt here is that if your soldiers are experiencing mental problems doing something then perhaps you should not be doing it.
I am not suggesting that the soldiers on the ground have much choice, but rather that arming soldiers with non-lethal weapons is perhaps not the right answer to the problem.
When people say "shouldn't we do something to stop it?", they really mean "shouldn't we give the government vast new expanded powers to regulate society, because only the government authority is efficient and trustworthy enough to solve the problem of pollution". The concept of massive government regulation and central-planning are implicit in what you are saying, because absolutly no-one of any political persuation wants to stop people from voluntarily acting to stop global warming.
Government does, and should, have a role in regulating markets. For instance, the government is responsible for labeling laws, and establishing standards measurements, and (more recently) in mandating industry use best common practices for accounting.
Even the most pro-HMV, anti-abortion, pro-gun, anti-public education Republican would not argue against this. The question is rather what the best role of the government is in the market.
One obvious way to influence the market is to apply the CAFE standards to all cars bought and sold, rather than exclude pickup trucks and SUV's. And in fact, the Bush administration has just last week proposed something like this, although of course in a very slight way designed not to upset major campaign donors at major car-building corporations.
The government can also shift spending away from projects that will encourage greenhouse emission, such as building new highways, to things like providing real alternatives to the car. This does not necessarily mean spending more money, but rather to spend it differently.
A third (and probably most important) way that the government can help is to fund basic (and applied) research to help minimise the demand for CO2-emitting fuel sources. Most likely this will mean research into nuclear power - cleaner, cheaper fission plants in the medium-term, and fusion plants in the long term. Government-funded research is necessary for technologies that have no hope to be profitable in a decade or two. Companies need to make money!
Presumably the hope with armoring vehicles is that you will make it harder for insurgents to get the weapons they need to attack you, and thereby reducing the number of attacks. (I have heard that in Iraq every home is legally allowed one automatic weapon.)
Ultimately the goal I suppose is to make attacks so expensive that they resort to participating in the existing government. The chances of that seem pretty limited without understanding and resolving their grievances (by which I do not mean negotiating with terrorists, but rather understanding the reasons that they have such a willing pool of recruits).
It depends on what you mean by "performance", doesn't it?
We have a database server that is limited by seek speed. For us, splitting our databases and adding more disks is the way forward (until we hit the next bottleneck). So in this case, yes adding more drives equals more performance.
But, if by performance you mean "throughput", then you are correct that adding more drives will help - to a point. At some point your controller won't be able to keep up with all those drives, and adding more will actually give you no additional benefit.
For instance, the Maxtor Atlas 15K II has a top transfer rate of over 90 Mbyte/sec. Even the highest-speed SCSI interface only has a throughput of 320 Mbyte/sec.
If the US government had a bad track record with the root domain servers, maybe more Americans would feel differently about this topic. But, let's face it: the US government is doing a fine job of managing them right now.
Well, I and K and M are outside of the US completely (at least, the organisations that run them are - because of anycasting the location of any given root name server is a tricky thing to pin down). Only E and G and H really fall under the US government.
So I guess you really don't know what you're talking about. Because if you did you'd know that when the root server system was created one of the goals was to have it outside of the control of any single entity, to provide redundancy not only in the physical and topological level, but also on the political level.
So actually, I guess you really are an arrogant American who doesn't understand.
Dude, I've been working with IPv6 for 7 years or so.
It's got built in equivalent of IPSEC. That alone would go a long way in improving most computing environments.
"built in equivalent of IPSEC"?? Huh? Rather, you mean the IPv6 standard requires that IPv6 implementations must have IPSEC, I am guessing. IPv6 with IPSEC is no more secure than IPv4 with IPSEC.
"Improved routing" refers to, among other things, route aggregation which reduces the size of routing tables which is helped by the simplified header which reduce router processing loads.
You are confusing two things here. IIRC, IPv4 checksum includes the TTL count, which means it has to be recalculated at every hop. This was fixed in IPv6. It's been a few years, but I think that this is what people normally refer to by "simplified headers".
Route aggregation, OTOH, is directly a result of address allocation policies. The hope is that because we can give "enough" addresses to each ISP, that any given network will only have to advertise a single route, thus minimising the number of routes that routers must maintain. This is a beneficial (and as yet unproven) side-effect... of having lots of addresses!
Someone with more networking knowledge can clarify why the IPv6 functions are much better than the IPv4 ones, where they may appear to overlap.
If you're talking about the socket API, in my experience there is no real advantage to the IPv6 functions. They exist so that you can manipulate IPv6 addresses, nothing more, nothing less.
Slashdot Mirror
We have a solution now, IPV4, that, if utilized properly, would alleviate the need for an additional IPV...X solution.
There are well over 6 billion people in the world. Ideally each person would be able to use an IP phone.
http://dev.mysql.com/doc/mysql/en/mysqldump.html
Quoth the page:
We use this and have never had any problems, but I guess we don't have these long running updates (at least not at 12:01 when our backup runs).
Sure, it's a script. Who cares? We used to backup by locking the database and copying the tables, but when we moved to InnoDB that didn't work any more. I timed it, and the mysqldump is just as fast to save and restore as the binary files when you use the "--tab" option.
Very insightful!
The question to ask a_greer2005 is, "Would you have voted against Bush if you knew about this?" My guess is the answer is "no", because this would have meant voting for Kerry.
OTOH, I am quite sure there are parents who would see the FBI's action as a good thing - even Democratic parents.
This is classic wedge politics. Basically, anything that has more of a chance to cause voters on the other side to defect than your own voters qualifies.
The rare events make for more drama. The news is about drama, not NEWS.
:)
If something is not a rare event, in what way is it news? I mean, what are you expecting?
ECONOMIC NEWS: Businesses Exchanging Goods and Services for Money
FASHION NEWS: Fashion Elite Wear Warmer Clothes in Winter than Summer
HEALTH NEWS: Ageing Linked to Death
SCIENCE NEWS: Earth Revolving Around Axis Once Per Day
News tends to be about things that don't happen every day. Except news about Paris Hilton's latest scandal, but some things never get old.
I am running on a 1.7 GHz laptop right now, and it takes less than one second to start up Konsole:Did you actually time the startup of the shell? Even if you did, do you really notice the 0.5 seconds of difference? For myself, KDE's "bloat" actually increases my efficiency, because I can create a new tab in Konsole rather than opening a new shell.
As for the desktop switching... the desktops switch faster than I can click - I tried to use the hotkeys to outrun the system (Ctrl-F1, Ctrl-F2,
KDE is not inherently slow.
Gosh, that's an ugly site. And hard to read too, with the background image, green-on-green font, and so on.
Why would someone who cares about attractive presentation use an application put together by these people?
Oh right, because they are producing scientific papers and have to use LaTeX. WYSIWYM is better than using notepad...
"Hey 35543334, come here for a second, will ya?"
:)
You are vastly overestimating the population of the Netherlands.
On a more serious note, the US started requiring a social security number (SSN, a taxpayer ID for the American pension scheme) for children claimed as "dependents" at a younger and younger age through the 1980's. This is part of the bizarrely complicated tax system in the US, but basically the government reduces your taxes for each child you take care of. The reason for requiring the SSN was to reduce people cheating by claiming they were supporting children who did not really exist. Supposedly it worked, but I can't be bothered to dig up statistics.
And the location of Turkey happens to be just outside the Russian border.
s/just outside the Russian/on the Soviet/
For example, apparently there's a new human mutation which adds an extra strand to the DNA chain, so you have 3 coming off the end instead of two.
I guess you're talking about chromosomes, not DNS strands. Everyone gets just the one DNA strand (unless you count mitochondrial DNA), right?
What you describe sounds more like a chromosomal aberration than a mutation to me, but I suppose it is possible. Do you have a URL?
I, for one, welcome our new --E or 3-E chromosomed overlords!
You're mistaken. When there is a threat, we mutate.
I am not a biologist, but I am pretty sure that while mutation does occur, it is rare and very few mutant genes survive to reproduce themselves even once. In general it is gene selection that causes most evolution rather than gene mutation.
By this I mean the case where a rare, but non-harmful, gene becomes useful due to a change in environment. In such a case the gene gains advantage over its alleles and becomes common.
At firms I worked with (telephony companies, usually), scheduled downtime is not included in downtime numbers. Of course, it depends on the SLA, but this is how it worked in the "5 nines" days of Ma Bell. 5 nines (99.999% uptime) was basically a myth. :)
Interesting. Experience tells us that eventually someone will develop the strategy of cheating against the other team mode players, thereby gaining even more of an advantage.
Ultimately the "team mode" question should be an unavoidable part of on-line playing. Do I collude with other players? If I do, how honest should I be?
We switched to InnoDB many years ago, because, frankly, MyISAM had locking problems. :)
We switched to 4.0 a couple of years ago, because we wanted the mysqldump that supported transactions and thereby allowed us to make a consistent backup without trickery.
BTW, InnoDB benched at about half the speed of MyISAM for us for a single query. However, we normally have concurrency, and the speed was only about 10% less in this case. And the worst-case performance was seconds rather than minutes, so it was a clear win.
We haven't seen any problems since we switched away from MyISAM (and had not with MyISAM for a long time either).
Indeed. I did some benchmarking of MySQL verus Microsoft SQL Server Express 2005 recently.
MSSQL works about 4x slower if you do not prepare queries properly. Even so, MySQL is significantly faster for the queries that matter in our application (but MSSQL is faster for certain other queries).
MySQL 4.1 and newer support prepared statements, if you really want them. (I'll benchmark that tonight, now that I think of it...)
Our update server gets about 14 updates per second. We've never had MySQL corruption on the box.
Both of our query servers get between 200 and 300 queries per second, in addition to mirroring the update server, and have never had MySQL corruption.
Our data is small (about 12 Gbyte), so maybe we're just lucky, but I think heavy activity is probably not the problem.
We *did* get some corruption once, about 4 years ago. At the time we were paying for MySQL support, pushed our files to the developers, and MySQL was fixed.
Are you sure it's not a bug in the application?
Unfortunately due to bad language design, finding signed/unsigned problems is often a subtle problem in C.
:)
Here is a phrack article on the topic.
Personally, I think the OpenBSD folks are doing things the hard way, by using an insecure language as the foundation of their work. That's the problem with C - you have to remember everything you learned over years of programming, all the time, or you risk making a mistake that can not only cause crashes, but ultimately compromise your entire machine, if not your entire network.
But having said that, I do use OpenSSH and occasionally OpenNTPd (on machines with interfaces that go up and down a lot).
In the eyes of most environmentalists, (or at least the environmental lobbying groups) nuclear power is more of a threat to the environment than global warming.
:)
Possibly, but there simply is no alternative. It might be possible to set up vast areas of desert or ocean as solar collection points, but that would not be 100% environmentally neutral.
An hour with the Internet and a calculator will convince you that even an extremely efficient biomass program cannot replace current fossil fuel consumptions. Renewable resources such as wind and hydroelectric power can help, as can reduced consumption, but in the end we need nuclear power.
Rather, we need nuclear power if we want to continue the standard of living of the small fraction of the world's people that live in the west. The world needs it even more if it wants all of humanity to live in safety and comfort. (Unfortunately this is not really a primary goal for religious people, or for self-interested or nationalistic people, but it should be!)
Some estimates are that fission would supply us with power for thousands of years. The main issues with fission are that you can build really scary weapons with the technology, and it produces waste that lasts effectively forever. These can be addressed, but I really believe that fusion power is the ultimate answer.
I've heard about how demoralizing it can be for our troops to be forced to kill children
The Nazi's developed gas as a method of killing Jews because of psychological problems appearing in the soldiers who were ordered to kill them.
I think the lesson learnt here is that if your soldiers are experiencing mental problems doing something then perhaps you should not be doing it.
I am not suggesting that the soldiers on the ground have much choice, but rather that arming soldiers with non-lethal weapons is perhaps not the right answer to the problem.
When people say "shouldn't we do something to stop it?", they really mean "shouldn't we give the government vast new expanded powers to regulate society, because only the government authority is efficient and trustworthy enough to solve the problem of pollution". The concept of massive government regulation and central-planning are implicit in what you are saying, because absolutly no-one of any political persuation wants to stop people from voluntarily acting to stop global warming.
Government does, and should, have a role in regulating markets. For instance, the government is responsible for labeling laws, and establishing standards measurements, and (more recently) in mandating industry use best common practices for accounting.
Even the most pro-HMV, anti-abortion, pro-gun, anti-public education Republican would not argue against this. The question is rather what the best role of the government is in the market.
One obvious way to influence the market is to apply the CAFE standards to all cars bought and sold, rather than exclude pickup trucks and SUV's. And in fact, the Bush administration has just last week proposed something like this, although of course in a very slight way designed not to upset major campaign donors at major car-building corporations.
The government can also shift spending away from projects that will encourage greenhouse emission, such as building new highways, to things like providing real alternatives to the car. This does not necessarily mean spending more money, but rather to spend it differently.
A third (and probably most important) way that the government can help is to fund basic (and applied) research to help minimise the demand for CO2-emitting fuel sources. Most likely this will mean research into nuclear power - cleaner, cheaper fission plants in the medium-term, and fusion plants in the long term. Government-funded research is necessary for technologies that have no hope to be profitable in a decade or two. Companies need to make money!
Presumably the hope with armoring vehicles is that you will make it harder for insurgents to get the weapons they need to attack you, and thereby reducing the number of attacks. (I have heard that in Iraq every home is legally allowed one automatic weapon.)
Ultimately the goal I suppose is to make attacks so expensive that they resort to participating in the existing government. The chances of that seem pretty limited without understanding and resolving their grievances (by which I do not mean negotiating with terrorists, but rather understanding the reasons that they have such a willing pool of recruits).
It depends on what you mean by "performance", doesn't it?
We have a database server that is limited by seek speed. For us, splitting our databases and adding more disks is the way forward (until we hit the next bottleneck). So in this case, yes adding more drives equals more performance.
But, if by performance you mean "throughput", then you are correct that adding more drives will help - to a point. At some point your controller won't be able to keep up with all those drives, and adding more will actually give you no additional benefit.
For instance, the Maxtor Atlas 15K II has a top transfer rate of over 90 Mbyte/sec. Even the highest-speed SCSI interface only has a throughput of 320 Mbyte/sec.
Alternately, we could relax a bit.
Well, I and K and M are outside of the US completely (at least, the organisations that run them are - because of anycasting the location of any given root name server is a tricky thing to pin down). Only E and G and H really fall under the US government.
So I guess you really don't know what you're talking about. Because if you did you'd know that when the root server system was created one of the goals was to have it outside of the control of any single entity, to provide redundancy not only in the physical and topological level, but also on the political level.
So actually, I guess you really are an arrogant American who doesn't understand.
Well, the people who discovered quantum mechanics didn't agree with you. This is known as the Copenhagen interpretation.
Wikipedia kicks ass, by the way. There is a whole article discussing alternate interpretations of quantum mechanics, if you don't happen to agree with Mr. Bohr and his colleagues.
Read up on IPv6 some time.
Dude, I've been working with IPv6 for 7 years or so.
It's got built in equivalent of IPSEC. That alone would go a long way in improving most computing environments.
"built in equivalent of IPSEC"?? Huh? Rather, you mean the IPv6 standard requires that IPv6 implementations must have IPSEC, I am guessing. IPv6 with IPSEC is no more secure than IPv4 with IPSEC.
"Improved routing" refers to, among other things, route aggregation which reduces the size of routing tables which is helped by the simplified header which reduce router processing loads.
You are confusing two things here. IIRC, IPv4 checksum includes the TTL count, which means it has to be recalculated at every hop. This was fixed in IPv6. It's been a few years, but I think that this is what people normally refer to by "simplified headers".
Route aggregation, OTOH, is directly a result of address allocation policies. The hope is that because we can give "enough" addresses to each ISP, that any given network will only have to advertise a single route, thus minimising the number of routes that routers must maintain. This is a beneficial (and as yet unproven) side-effect... of having lots of addresses!
Someone with more networking knowledge can clarify why the IPv6 functions are much better than the IPv4 ones, where they may appear to overlap.
If you're talking about the socket API, in my experience there is no real advantage to the IPv6 functions. They exist so that you can manipulate IPv6 addresses, nothing more, nothing less.