Too many times I've been behind someone who was obviously talking on their mobile. I pass them (they're driving well under the speed limit because they're obviously unable to pay proper attention to all aspects of driving at the same time), fully expecting to see them holding it to their ear as I peer over.
They're not holding a phone to their ear, but they are talking to someone over hands free. They are distracted, no doubt about that. It's good potential for hilarity just watching people nearly bumping into things while talking as they walk. I've noticed that those who are good at going aroung obstacles tend to have their eyes stuck wide open in some kind of automatic avoidance trance, or to slow down, or even stop.
I'm glad that the one concession people do tend to make to loss of concentration is a lower speed. But puh-lease folks, try to keep talking on a mobile phone in the car to making urgent calls, and keeping incoming social calls quick (or if it's going to take a while, and you have the option, swap drivers).
And whatever you do... never... EVER... let a drunk elderly person make social calls from the car while they're driving!
Sweet spot? In Sydney and Melbourne, almost certainly. Perth however is 1.2 million or so people spread out over a couple of hundred square kilometres - and that's where iiNet was born. Adelaide's density wouldn't be much higher. Canberra, while designed well from conception is as a result mostly suburban for housing. And that's leaving out a few... submarine trunk to Tasmania anyone?
I believe many Australians are still dreaming of 180+ square metre homes on quarter acre blocks, and even if they don't get the desired space, that's keeping a large area of the lower density capital cities from building up instead of out.
Telcos are overcharging us (Telstra has to look after its shareholders after all), but they are also still installing infrastructure. It's fine installing the infrastructure for a city, but running a trunk over 3000km to the next nearest core hub is never going to be a cheap exercise, especially when we still get the cost of backhoe related outages adding to it.
Plus they have to keep upgrading capacity because people keep demanding more! And getting service to population centres off the beaten track does matter to the customers, and that adds to the cost, which is shared by all, whether we believe it's fair or not.
I find it refreshing to see ISPs installing their own infrastructure, but the costs are still there. At least call costs have generally come down since privatisation of the telco industry.
I'm still curious as to whether/how much Australian ISPs are still being charged by upstream (U.S.?) providers for volume. That certainly wouldn't help, though obviously those costs have at least gone down. Anyone?
Being such a geographically sparse country, investment in infrastructure generally involves a lot more up-front cost than say, Sweden (most European countries come to think of it), per person. Think about it - driving approximately around the mainland coast would take about 10 days at a guess if you were pushing it, and there's only around 20 million people on it. That's a lot of man hours, cables and equipment to install. So don't expect cheap all-you-can-eat access in Australia yet.
There are various plans at various rates - one provider offers 512/128 for $70/month with no restrictions, not sure about the cost for higher peaks. I wouldn't look for any vast improvement over this sort of capped plan for at least another 5 years, and that's only assuming the standards don't improve the peak speed even further.
iiNet have spent $10M on installations, and only have customers numbering in the tens of thousands of dollars. They obviously can't give the service away, but the rates are still reasonable especially compared with the telco offerings. As I understand it, there are still per Mb costs from at least some of our international trunk providers too. Anyone who can refute that, or that has details?
Given that 90-95% of computers run some version of Microsoft Windows, that other 5-10% appear to be rather worse at spreading spam, eh?
RTFP... 'e said that Zombie Windows boxes contribute 80% of spam, and that implies the bulk of the rest are due to abuse of services which people have otherwise legitimate access to - SPAMmers' own boxes - some probably running Linux because the OS is free and they'd tackle the job of sending SPAM very effectively, but nothing stopping SPAM software running on any platform in particular because it's your box!
I'll not say anything for or against the percentage stated, but that level certainly wouldn't surprise me. I am however sorry to say that there are probably a few open relays still out there in *nix world. This is very different to being owned by a hacker/script kiddie, but it is a result of poor (not insecure!) configuration.
You'll probably find that something like 80% of SPAM comes from Zombied Windows PCs, and almost of all the rest from the SPAMmers own PCs, whether they run Windows, Linux or OS X. Maybe a very small fraction of a percent would be poorly administered Linux/Unix systems that have been rootkitted for the purpose. It's generally just too hard to bother trying (check out the links at the bottom of the article).
Unfortunately that's all too true. Blame the leader, not the follower. I'd love to see legislation counter this by pushing companies to at least have partial accountability. Something along the lines of:
"If our software screws up your business, we must try to fix it. You must give us X days from the initial report to stop/fix the problem. If we can't do that, then you may sue for loss of productivity".
The value of X should be inversely proportional to the license cost of the software, before consideration of any maintenance contract. In essence, enforce an accountability relationship on the software vendor based on cost - the sort of response times FOSS authors give of their own accord. Of course, this type of clause should expire once the product is past its use by date, but it might make them care more about security and reliability.
Would dodgy software vendors stay afloat and risk being sued into bankruptcy? I imagine, for example, Microsoft's OS business might struggle a bit, though they'd still be making money from the Office suite while people continue to buy it. Or would they be forced to lower the cost of their less reliable products to meet their response times to real problems? Or would they actually get their acts together and show genuine concern for their products? And how well would RedHat deal with this? (Pretty well, I would expect, though they would be no less vulnerable than the competition). I wonder...
"My" thinkpad uses the touchpad as the "mouse wheel"... vertical AND horizontal. Lotsa fun! I rarely use the middle button that's provided. Not sure about with Linux, but (Windoze)...
Just drag your finger along the right hand side for scroll up/down or along the bottom for left/right. Speed of drag affects speed of scroll, and if you lose contact while you're moving your finger, it stays scrolling at that speed.
If you're in a touchpad only situation, no problem. If not, it can cause confusion - say you used it to set a scroll speed, and it finally scrolled all the way down to the bottom of a document and then you used a mouse to scroll back up to the top - the touchpad scroll pushes it back down again at the original speed. In any window that may happen to get focus. Potentially confusing 'til the penny drops. Fun for all the family!
Hmmm... I can think of one way. Current voice recognition technology may pick enough words out to make the call indexable (language dependent?) - and helpfully store your voice calls for later perusal. Depends on how much they want to get into your life, and how much you want to let them.
Yeah, the RAID/md metadata lives at the end of the partitions, which boot loaders generally don't look at. And indeed, LILO has its uses (installing onto hda1 and hdc1 instead of md1 or whatever without disrupting the mirror)
Of course, again, if you have the luxury of good hardware RAID, then you don't even have to think about this.
If Werner is going to claim copyright, he should state his sources - there is very little chance that he wrote every word.
There is every chance, actually. I've read some of Werner's Oracle installation stuff - it's very much "what I learnt bashing my head against all these problems" digested relatively neatly into a pseudo-howto. From what I've seen he generally writes from experience (or his notes), and hiss mis-steaks are hiss onw.;)
Good Stuff Werner. I wish I had this when I was setting up our last RAC cluster - would have saved me many hours!
While I'm sure with enough fiddling it can be done,/boot generally doesn't like living in an LVM filesystem. The main reason for that would be the boot loader needs to load extra and interpret the metadata to figure out where the initrd you want is loaded from. I imagine most distros simply won't let you put/boot on a logical volume because it requires too much extra.
I did manage to get/boot to live happily on a software RAID1 mirrored partition - this was based on a RH 7.3 build, but RedHat won't do that for you by default (do any?).
Of course, it's not such a problem with proper enterprise storage (but you still can't put boot on an LV, as a rule)...
Hmmm... the drop-off would also correlate reasonably well with the destruction caused by the Tsunami, and how long it would have taken in some cases to get the infrastructure going again(probably more getting power back than network structure). My incoming SPAM dropped about 80% or more for about a week.
Indeed. Let's not forget that now we live in a world where everyone else can be held accountable for your actions unless they can prove that they tried to warn you. Ignorance of what should come completely naturally for survival is tolerated, and in some cases encouraged. We tend to focus on "me" and maybe people close to us, ignoring the needs of others, and the environment. Accountability is someone else's problem.
I believe it's not really survival of the fittest any more, which includes fitness of instincts, so much as how effectively your support network runs, or how well you milk it - using any support from hand-outs and social welfare (the system is accountable for making me thrive), tools, research, and common sense (I will hold myself accountable for my survival), all the way to utilising advisors and bodyguards (others are accountable for my health/safety)...
If left to their own devices, what proportion of each of these generalised groups will go collecting pretty shells and flapping fish when the tide suddenly goes out?
Hmmm... I would put it down to fish generally not being particularly bright outside of their day-to-day survival instincts, and as they are generally spread out, not needing to respond to such major events as a tsunami for the species to go on. They'd stay in their usual habitat, and if that local habitat goes, then the inhabitants go with it. Occasionally they'd ride a current, especially if given no choice. If they get washed ashore, bad luck.
What I'd be more interested in is if any, and if so how many, larger sea creatures were dumped ashore. (I believe there were whale beachings in the general vicinity within that general window, but AFAIK it's just coincidental).
I'm no marine biologist, so anyone with facts that support or shoot down my ideas are quite welcome to do so.
Hmmm looks like you started using Linux about the same time I did.:) I have been running my own (or someone else's) system since the 0.92 or possibly 0.96 days. I've forgotten a lot about Linux too, including a detail or two on my first install.
We are dealing with ES3 here. A simple case of new(ish) architecture generated two support calls which I resolved before Dell could find someone to get the answers off, and that includes one or two chats to a RedHat engineer. Dell didn't know much about it (rather than try to find out what was wrong with the software, every component except the case and CPU were replaced). RedHat was able to get back to us after a couple of days and a couple of panic dumps in e-mail. I wish we would more consistently get the experts instead of the first and second level (we have to do that through Dell) going around in circles, but failing that the escalation should be automatic.
Yes, Oracle 10g is certified and works quite happily with ES3 - has since it was released. The same couldn't be said of Oracle's iLearning product working with 10g on ES3 until recently, or their OID product working with the fabulous RAC backend storing the database (recently certified). Just a few months ago, for iLearning to install, their compatability matrix forced us to use AS2.1 (Update 5) on our (IBM) blades that would install onto but not boot off without a hand compiled LILO upgrade. All for one feature, an out of the box install was suddenly required to happen on older hardware. This was not IMHO RedHat's fault.
I could bitch about a few things RedHat have decided (not) to do, there are usually valid reasons behind them and it's already gone through channels to them. I don't expect RedHat to change their minds, but at least they've made their position on them clear. However, at the end of the day, the Enterprise Linux product is certainly a highly professional release.
If I weren't now rusty with PostgreSQL (it's been a few years), and still had an up-to-date manual, I would have no problems supporting it myself at all bar the code level (used to play with the guts more than the user side for a living in the 7.2 days), and I would bet that I would now be far more likely to get efficient support from PostgreSQL support organisations within hours than from a top priority RedHat support call. Why? Because they're not burdened by their support chains, and because PostgreSQL is really just a single entity, with no complex relationships, and from my experience everything you need is in one of the manuals. It's a shame that some of these will grow to be corporate entities bogged down by self-imposed red tape. I do wish them luck avoiding it.
The latest we've heard where I work is that to use the EM64T processor (pushed as a package solution by Intel, Oracle, Dell and RedHat), we need RedHat's ia32e kernel (as opposed to ia32). I hope this is right. Please tell me if it isn't so we can avoid wasting more time!
you need 24/7 support and someone who knows both the software and hardware/OS intimatly
Funny, I'm yet to deal directly with anyone from Oracle who knows the OS that well at all. In this case, Linux. From my experience, big companies with impressively high priced support contracts tend to get back to us quickly, sober, in front of a computer and with no screaming wife/kids in the background. If the problem is mainstream, or we're just plain lucky, then the answer/solution comes quickly. But that's about the extent of it. With our team expertise we usually get a better response from a well phrased Google search.
And it's not just Oracle. It's also Dell, and RedHat. Can Oracle really run in 64 bit mode on an Intel EM64T processor? Does RedHat's 3rd level support know which kernel makes the most of that architecture? Can Dell's 3rd level guys tell us what is required for this, or what is normal? From my experience, the answer is usually: Not before I stumble across the answer on Google.
I don't care how many people staff the company that supports us. Effective staffing of any support oriented company comes down to quality. The small business that knows your application inside out is your friend. If the guy on the phone has been playing in the CVS tree then you can usually expect a quick and correct answer or bug fix. Not that I expect to encounter any bugs in any stable OSS release.
show me something in the open-source world that can do 1000+ transactions per second, with complete atomicity, and ability to pull the plug on that system and then seamlessly roll it back to the exact moment in time that it was at when it died...
I can. Did it 1 or 2 years ago with PostgreSQL 7.something on a Linux box, on a Reiser 3.6 filesystem with patches to enable full journalling on the database filesystem. I treated the machine like an old fashioned DOS box and simply turned it off when I felt like it. Short of hardware failure, I could not kill it. In the boot that followed it committed all the disk activity it could and the service simply started with the last (partial) transaction cleanly rolled back.
At the same time, I could not kill the Oracle database I installed on the same filesystem (though it only took 1 interruption to render it corrupt on ext3 with only metadata journalling). The only difference then is what happens when your admin screws up and blows away important files and you need to restore from backups.
Of course, Oracle has had that under control for quite some time - you can roll back/restore data to whatever point in time you like if you have taken sufficient backups, redo and archive logs. Now PostgreSQL has just introduced something to provide similar ability (I want to see revised manuals!), and appears to be a real option. I'm itching to play with PostgreSQL 8 on a Reiser 4 filesystem if I get the chance! (I believe still on Reiser 4's TODO list is an API to turn an arbitrary set of disk operations into a completely atomic operation. This IS funky!)
It's hard to go past how well Oracle RAC clusters can tolerate hardware catastrophes, but I still have a gripe with it, not so much in the backend as all the Java apps that come with it - yes it appears they were written to run on anything, but the memory sacrifice is far from insignificant. And as for support, well... just make sure the Oracle app is certified to run with your chosen backend!
PostgreSQL on a reliable filesystem with suitable mirroring/striping is simply a rock solid alternative to the commercial platform. It works - reliable and fast. MySQL is yet to particularly impress me (it always seems to be lagging behind the leaders), though it is continually improving.
Who can speak for the others? Open sourcing might be the right thing to do, but the competition is fierce. Hopefully they will remain viable to keep the game more interesting.
Oracle still has a few enterprise features that most of the open source databases lack-- table partitioning,...
At least now with PostgreSQL 8 tablespace partitions become a reality. Plus point-in-time backup and recovery, another valuable feature. I'm unaware of any other F/OSS databases that can do this. Clustering natively, well maybe in PostgreSQL 9?
I found it refreshing that although it was possible, quite a bit of work had to go into hacking the bikes in the first place. The only way to circumvent the system was to physically get inside it and reprogram the firmware. This is solid engineering IMHO.
Now, if they're really recalled over winter for a once-over, I wonder how much harder they will be to get into when the snow melts and they reappear? Will the Call-A-Bikes become more impenetrable, as Darwinian geek theory dictates they'll need for survival?
On the flip side, I must also applaud the hackers for not simply vandalising the bikes, but modifying them to return to service once the "free ride" was over. (Perhaps misguided) - theft all the same, but at least free of physical destruction.
Well, actually, why not sign the software themselves? Think about it. The author of the article was having an issue with the apparently "random" nature of the location of the download mirror.
Even if it's not signed through an official CA, at least they can follow it back to mozilla.org from W.X.Y.Z - and then making a trojan out of the software would require successful simultaneous hacking of 2 sites.
I hope these things stay up there and prove viable. Naturally, the closer to the ground the s(tr)atellite is, the better the response time. That's always been the problem with traditional geosync satellites - while they could usually handle high bandwidth by allowing large packet windows, there was always significant lag, which kinda interferes with the online gaming.
Will we stratellites in place as an alternative for when backhoe incidents occur? And if so, when?
Slashdot Mirror
They're not holding a phone to their ear, but they are talking to someone over hands free. They are distracted, no doubt about that. It's good potential for hilarity just watching people nearly bumping into things while talking as they walk. I've noticed that those who are good at going aroung obstacles tend to have their eyes stuck wide open in some kind of automatic avoidance trance, or to slow down, or even stop.
I'm glad that the one concession people do tend to make to loss of concentration is a lower speed. But puh-lease folks, try to keep talking on a mobile phone in the car to making urgent calls, and keeping incoming social calls quick (or if it's going to take a while, and you have the option, swap drivers).
And whatever you do... never... EVER... let a drunk elderly person make social calls from the car while they're driving!
I believe many Australians are still dreaming of 180+ square metre homes on quarter acre blocks, and even if they don't get the desired space, that's keeping a large area of the lower density capital cities from building up instead of out.
Telcos are overcharging us (Telstra has to look after its shareholders after all), but they are also still installing infrastructure. It's fine installing the infrastructure for a city, but running a trunk over 3000km to the next nearest core hub is never going to be a cheap exercise, especially when we still get the cost of backhoe related outages adding to it.
Plus they have to keep upgrading capacity because people keep demanding more! And getting service to population centres off the beaten track does matter to the customers, and that adds to the cost, which is shared by all, whether we believe it's fair or not.
I find it refreshing to see ISPs installing their own infrastructure, but the costs are still there. At least call costs have generally come down since privatisation of the telco industry.
I'm still curious as to whether/how much Australian ISPs are still being charged by upstream (U.S.?) providers for volume. That certainly wouldn't help, though obviously those costs have at least gone down. Anyone?
Must have misplaced my marbles.
There are various plans at various rates - one provider offers 512/128 for $70/month with no restrictions, not sure about the cost for higher peaks. I wouldn't look for any vast improvement over this sort of capped plan for at least another 5 years, and that's only assuming the standards don't improve the peak speed even further.
iiNet have spent $10M on installations, and only have customers numbering in the tens of thousands of dollars. They obviously can't give the service away, but the rates are still reasonable especially compared with the telco offerings. As I understand it, there are still per Mb costs from at least some of our international trunk providers too. Anyone who can refute that, or that has details?
Given that 90-95% of computers run some version of Microsoft Windows, that other 5-10% appear to be rather worse at spreading spam, eh?
RTFP... 'e said that Zombie Windows boxes contribute 80% of spam, and that implies the bulk of the rest are due to abuse of services which people have otherwise legitimate access to - SPAMmers' own boxes - some probably running Linux because the OS is free and they'd tackle the job of sending SPAM very effectively, but nothing stopping SPAM software running on any platform in particular because it's your box!
I'll not say anything for or against the percentage stated, but that level certainly wouldn't surprise me. I am however sorry to say that there are probably a few open relays still out there in *nix world. This is very different to being owned by a hacker/script kiddie, but it is a result of poor (not insecure!) configuration.
You'll probably find that something like 80% of SPAM comes from Zombied Windows PCs, and almost of all the rest from the SPAMmers own PCs, whether they run Windows, Linux or OS X. Maybe a very small fraction of a percent would be poorly administered Linux/Unix systems that have been rootkitted for the purpose. It's generally just too hard to bother trying (check out the links at the bottom of the article).
Seriously, generalisations aside (both in the original, and in the more accurate versions), the original is much harder to believe for some reason.
... for this black duck ... is the list of related links at the bottom of the page. Sometimes the best chuckles come from RTFA!
"If our software screws up your business, we must try to fix it. You must give us X days from the initial report to stop/fix the problem. If we can't do that, then you may sue for loss of productivity".
The value of X should be inversely proportional to the license cost of the software, before consideration of any maintenance contract. In essence, enforce an accountability relationship on the software vendor based on cost - the sort of response times FOSS authors give of their own accord. Of course, this type of clause should expire once the product is past its use by date, but it might make them care more about security and reliability.
Would dodgy software vendors stay afloat and risk being sued into bankruptcy? I imagine, for example, Microsoft's OS business might struggle a bit, though they'd still be making money from the Office suite while people continue to buy it. Or would they be forced to lower the cost of their less reliable products to meet their response times to real problems? Or would they actually get their acts together and show genuine concern for their products? And how well would RedHat deal with this? (Pretty well, I would expect, though they would be no less vulnerable than the competition). I wonder...
Just drag your finger along the right hand side for scroll up/down or along the bottom for left/right. Speed of drag affects speed of scroll, and if you lose contact while you're moving your finger, it stays scrolling at that speed.
If you're in a touchpad only situation, no problem. If not, it can cause confusion - say you used it to set a scroll speed, and it finally scrolled all the way down to the bottom of a document and then you used a mouse to scroll back up to the top - the touchpad scroll pushes it back down again at the original speed. In any window that may happen to get focus. Potentially confusing 'til the penny drops. Fun for all the family!
Nay-sayers, feel free to say nay...
Of course, again, if you have the luxury of good hardware RAID, then you don't even have to think about this.
There is every chance, actually. I've read some of Werner's Oracle installation stuff - it's very much "what I learnt bashing my head against all these problems" digested relatively neatly into a pseudo-howto. From what I've seen he generally writes from experience (or his notes), and hiss mis-steaks are hiss onw. ;)
Good Stuff Werner. I wish I had this when I was setting up our last RAC cluster - would have saved me many hours!
While I'm sure with enough fiddling it can be done, /boot generally doesn't like living in an LVM filesystem. The main reason for that would be the boot loader needs to load extra and interpret the metadata to figure out where the initrd you want is loaded from. I imagine most distros simply won't let you put /boot on a logical volume because it requires too much extra.
I did manage to get /boot to live happily on a software RAID1 mirrored partition - this was based on a RH 7.3 build, but RedHat won't do that for you by default (do any?).
Of course, it's not such a problem with proper enterprise storage (but you still can't put boot on an LV, as a rule)...
Any concrete figures?
I believe it's not really survival of the fittest any more, which includes fitness of instincts, so much as how effectively your support network runs, or how well you milk it - using any support from hand-outs and social welfare (the system is accountable for making me thrive), tools, research, and common sense (I will hold myself accountable for my survival), all the way to utilising advisors and bodyguards (others are accountable for my health/safety)...
If left to their own devices, what proportion of each of these generalised groups will go collecting pretty shells and flapping fish when the tide suddenly goes out?
What I'd be more interested in is if any, and if so how many, larger sea creatures were dumped ashore. (I believe there were whale beachings in the general vicinity within that general window, but AFAIK it's just coincidental).
I'm no marine biologist, so anyone with facts that support or shoot down my ideas are quite welcome to do so.
Oh, that is so yesterday! Keep up, people!
We are dealing with ES3 here. A simple case of new(ish) architecture generated two support calls which I resolved before Dell could find someone to get the answers off, and that includes one or two chats to a RedHat engineer. Dell didn't know much about it (rather than try to find out what was wrong with the software, every component except the case and CPU were replaced). RedHat was able to get back to us after a couple of days and a couple of panic dumps in e-mail. I wish we would more consistently get the experts instead of the first and second level (we have to do that through Dell) going around in circles, but failing that the escalation should be automatic.
Yes, Oracle 10g is certified and works quite happily with ES3 - has since it was released. The same couldn't be said of Oracle's iLearning product working with 10g on ES3 until recently, or their OID product working with the fabulous RAC backend storing the database (recently certified). Just a few months ago, for iLearning to install, their compatability matrix forced us to use AS2.1 (Update 5) on our (IBM) blades that would install onto but not boot off without a hand compiled LILO upgrade. All for one feature, an out of the box install was suddenly required to happen on older hardware. This was not IMHO RedHat's fault.
I could bitch about a few things RedHat have decided (not) to do, there are usually valid reasons behind them and it's already gone through channels to them. I don't expect RedHat to change their minds, but at least they've made their position on them clear. However, at the end of the day, the Enterprise Linux product is certainly a highly professional release.
If I weren't now rusty with PostgreSQL (it's been a few years), and still had an up-to-date manual, I would have no problems supporting it myself at all bar the code level (used to play with the guts more than the user side for a living in the 7.2 days), and I would bet that I would now be far more likely to get efficient support from PostgreSQL support organisations within hours than from a top priority RedHat support call. Why? Because they're not burdened by their support chains, and because PostgreSQL is really just a single entity, with no complex relationships, and from my experience everything you need is in one of the manuals. It's a shame that some of these will grow to be corporate entities bogged down by self-imposed red tape. I do wish them luck avoiding it.
The latest we've heard where I work is that to use the EM64T processor (pushed as a package solution by Intel, Oracle, Dell and RedHat), we need RedHat's ia32e kernel (as opposed to ia32). I hope this is right. Please tell me if it isn't so we can avoid wasting more time!
Funny, I'm yet to deal directly with anyone from Oracle who knows the OS that well at all. In this case, Linux. From my experience, big companies with impressively high priced support contracts tend to get back to us quickly, sober, in front of a computer and with no screaming wife/kids in the background. If the problem is mainstream, or we're just plain lucky, then the answer/solution comes quickly. But that's about the extent of it. With our team expertise we usually get a better response from a well phrased Google search.
And it's not just Oracle. It's also Dell, and RedHat. Can Oracle really run in 64 bit mode on an Intel EM64T processor? Does RedHat's 3rd level support know which kernel makes the most of that architecture? Can Dell's 3rd level guys tell us what is required for this, or what is normal? From my experience, the answer is usually: Not before I stumble across the answer on Google.
I don't care how many people staff the company that supports us. Effective staffing of any support oriented company comes down to quality. The small business that knows your application inside out is your friend. If the guy on the phone has been playing in the CVS tree then you can usually expect a quick and correct answer or bug fix. Not that I expect to encounter any bugs in any stable OSS release.
I can. Did it 1 or 2 years ago with PostgreSQL 7.something on a Linux box, on a Reiser 3.6 filesystem with patches to enable full journalling on the database filesystem. I treated the machine like an old fashioned DOS box and simply turned it off when I felt like it. Short of hardware failure, I could not kill it. In the boot that followed it committed all the disk activity it could and the service simply started with the last (partial) transaction cleanly rolled back.
At the same time, I could not kill the Oracle database I installed on the same filesystem (though it only took 1 interruption to render it corrupt on ext3 with only metadata journalling). The only difference then is what happens when your admin screws up and blows away important files and you need to restore from backups.
Of course, Oracle has had that under control for quite some time - you can roll back/restore data to whatever point in time you like if you have taken sufficient backups, redo and archive logs. Now PostgreSQL has just introduced something to provide similar ability (I want to see revised manuals!), and appears to be a real option. I'm itching to play with PostgreSQL 8 on a Reiser 4 filesystem if I get the chance! (I believe still on Reiser 4's TODO list is an API to turn an arbitrary set of disk operations into a completely atomic operation. This IS funky!)
It's hard to go past how well Oracle RAC clusters can tolerate hardware catastrophes, but I still have a gripe with it, not so much in the backend as all the Java apps that come with it - yes it appears they were written to run on anything, but the memory sacrifice is far from insignificant. And as for support, well... just make sure the Oracle app is certified to run with your chosen backend!
PostgreSQL on a reliable filesystem with suitable mirroring/striping is simply a rock solid alternative to the commercial platform. It works - reliable and fast. MySQL is yet to particularly impress me (it always seems to be lagging behind the leaders), though it is continually improving.
Who can speak for the others? Open sourcing might be the right thing to do, but the competition is fierce. Hopefully they will remain viable to keep the game more interesting.
At least now with PostgreSQL 8 tablespace partitions become a reality. Plus point-in-time backup and recovery, another valuable feature. I'm unaware of any other F/OSS databases that can do this. Clustering natively, well maybe in PostgreSQL 9?
Now, if they're really recalled over winter for a once-over, I wonder how much harder they will be to get into when the snow melts and they reappear? Will the Call-A-Bikes become more impenetrable, as Darwinian geek theory dictates they'll need for survival?
On the flip side, I must also applaud the hackers for not simply vandalising the bikes, but modifying them to return to service once the "free ride" was over. (Perhaps misguided) - theft all the same, but at least free of physical destruction.
Even if it's not signed through an official CA, at least they can follow it back to mozilla.org from W.X.Y.Z - and then making a trojan out of the software would require successful simultaneous hacking of 2 sites.
Will we stratellites in place as an alternative for when backhoe incidents occur? And if so, when?
More targets to avoid the pressure of /.ing?