I suppose what the article means is that there are 4 x 48U racks installed in the server room. It is fiction that each rack could actually loaded with 48 x 1U servers! Potential problems are: cooling, weight, air (fire hazard), power supply.
Most likely actual rack usage looks as follows:
- Rack with 5 Servers - Rack for Patching and switches - Rack for phone system / phone patches - Rack for backup.
If they have remaining capacity, they could rent it out/sell to other community organisations.
Sound to me like the poster uses a virtual shared server, ie. The hosting company provides a virtual root environment, this would be an explanation why they can "root" the server, ie. just access the file system directly from the hypervisor.
I am sure this kind of service does not include uptime guarantees, it operates most likely under "Best Effort" promise. Which means they do not need to guarantee either uptime or availability. Their monitoring system and logs do not detect any error, so they want to check the posters systems logs - a reasonable request as they are trying to help him.
Monitoring is a difficult task - what to you measure, where is your sensor. Was the outage on your own internet connection rather than on the server?
"Moments ago, there were three simultaneous outages while I was logged into the server working on some projects. " It sounds very suspicious, like the admin interface causes interruption in the VM. Something like this happened to my VMWare Servers totally unexplainable, but reproducably.
Its a silly argument that the article makes. E-Readers can display B&W pages, and gray shades. This is absolutely sufficient for displaying textbooks.
There is a line of extremely bloated textbooks that takes liberties with layout and colour. I just wonder for what purpose colour is used in those books.
In my opinion it does neither help understanding of diagrams and models, nor does it improve the information density.
Perhaps a revisit of Tufte's rules for information design is in order for publishers?
Companies operate with constrained resources in order to generate a profit. While developing a product a company only experiences cost, no income. After the selling starts a company receives income that slowly covers development cost until it breaks even. After break even, additional sales generate profit. After sales start any maintenance work reduces profit.
Why do I talk about what everyone already knows? A company like microsoft must decide for themselves how to use their limited resources. They decide how serious a bug is by looking at the urgency, and seriousness vs reduced profits. They make a decision based on their own interest. By putting an exploit on ebay, the cost and seriousness of the bug is not decided merely by microsoft, but by the market.
If MS thinks its irrelevant, they need not do anything. If they think its serious, they will have the choice of either fixing it until the auction ends, or bidding to prevent disclosure. Third parties interested in security will also have a chance to bid, however they must realise that their purchase may soon loose value if a fix is provided.
Overall I think its a good idea - in the best case it encourages a faster fixing of issues.
I agree. Converted from NT4 Server to Samba. There are some imperfection with oplocks, but samba itself works reliable in an NT Domain or Workgroup - after all Mac OS/x also uses Samba, and the author said that it has no problem!
Suse often supplies RPMs that do not work properly are configured in a strange way. If you are an ordinary customer they will not tell you the little secret that makes it work.
Saying that though, it should still be possible to make Suse's samba work. I suppose a configuration wizard would help this kind of user.
But to be fair, they tested the office version, so you would expect the essentials business apps to work straight away without tinkering.
Well seems that the guy who forked gentoo with his "great" gentoo embedded idea does not have much of a business proposition after all. And why should he, with competition from QNS and existing Linux embedded projects.
First of all, if I remember correctly Gallileo was conceived a long time ago, way before recent tensions between the US and Europe. It addressed the problems with the availability and accuracy of the American GPS system, so that it can be used (mainly) in Europe for civilian purposes: Automatic Freight unloading and transhipment in Ports, Aircraft & Ship navigation, as well as many more commercial activities.
What is the age of GPS satelites, and what is their useful working life? Are they really still state of the art? I think with Gallileo, the world has the chance to obtain a newer and better technology.
No matter what your political opinion, we should all welcome the better system. Some People are happy to run Suse or RedHat with low performance, others prefer optimised Gentoo. When the system is ready, you can still continue to use GPS if you want to. Nobody is forced to switch.
They are handmade, with a three year warranty and about as good as it gets.
In my company we have hundreds of them and they never fail (both LCD & CRT).
"Someone" I know, just downloaded Freelancer - a Microsoft Game - and is really happy with it. He was thinking of buying the boxed version anyway, to get the a nice manual, and additional features like a printed starmap etc...
He then discovered on the MS Website, that the boxed version actually does NOT contain a manual!
Of course, now there is no reason to buy it anymore.
My conclusion is that people download warez games because of two reasons:
- They cannot get it in the shop (my friend who lives in Japan) - There is less value in the product, but at the same price(Privateer and Wingcommander Prophecy had both Manuals and consisted of several CDs!!)
I guess your bad experience gives you the right to be scared of two wheel vehicles. But does it mean just because something is more dangerous than, lets say walking, one shouldn't do it anymore?
Knowing about the risks, one should perhaps decide on strategies to avoid the main problems: Avoid the main roads, rush hour, dangerous corners, etc...
Here in Tokyo I started to commute to work. I avoid the major ateries leading into the city centre, instead use small roads through residential areas & small paths along rivers. In very bad places I cycle on the sidewalk. On certain small roads with two way traffic, I block cars behind me from overtaking in dangerous places by cycling in the middle of the road.
Drivers here are so bad that, you may just as likely be hit walking! But then I suppose in LA, in addition to the dangers of traffic, you could also get shot if you go through the wrong neighbourhood?
Just purchased a road type bicycle on Friday. Wow, have things changed. The last time I rode a half decent bike was 15 years ago.
It is all high tech mechanics nowadays: Aluminium Alloy light weight frame, Gear Levers integrated into the breaks, High pressure tires.
As a result the thing accelerateslike hell even bearing my unfit body. Adrenalin rushes to your head, while you zip in and out of the traffic, overtaking traffic jammed cars - you are the true king of the road!
How can a mere home training bike compete?
How about giving M$ a taste of their own medicine? NT users have to pay per server and per license fee, so why not charge charge M$ per square meter and per pedestrian that walked by at the time?
There is no point crying about all the tiny features that don't work in Suse. First of all, each distribution has a different way of doing things, which is nice as it shows that we have freedom of choice.
So when reviewing a distribution you need to look at it objectively from the value it gives you. People who have been using yast or yast 2 before are very effective in using it and working around the querks, just like many Linux user are beter with the bash or bourne shells than Windows users.
Secondly, most people are not interested in hearing the same old problem stories, after the authors keeps throwing all sorts of weird and exotic hardware-software combinations at the distro. For my part I am interested how well it works on normal PCs, eg. Dell Optiplex, Poweredge, Compaq Proliant servers, etc... I am sure in Germany (where Suse is located) there are many strange monitors that are 2 years old that are not auto-detected by RedHat et al! So saying that a particular monitor is not on the hardware list constitutes hardware compatibility issues is pretty daft(the same could be said about any linux system).
Also - graphical user interfaces are purely a matter of personal taste. If you don't like KDE, please use Gnome. It is no secret that Suse is KDE biased, so please do not blame them if you don't like the looks. It cannot be that difficult to download a theme or customize the desktop to fit your private preferences?
So at the end of the day, what value does Suse add?
- Grub substitutes Lilo, a necessary evolutionary step - CUPS printing as default brings it up to the current state of the art - XFS file system introduced already in 8.0 brings reliable journaling and ACLs for those people who do not like patching kernels and using ReiserFS or EXT3 - Apache, Samba, Bind and other standard apps precompiled with workable configurations
If you are a pro, you do not have to use these SUSE RPMs or yast, but instead you will compile these servers yourself anyway.
Overall Suse is not perfect, but neither are the other distros. Most problems are a matter of personal preference, but these problems also happen with other distros. The point is that Suse gets the job done, and once you take of your red hat and learn doing it the other way, you can actually get a nice result.
Well actually there is no such thing as a single 3G standard! As far as I know there are several competing standards mainly the European one and the American one (by Qualcom).
As a matter of fact here in Japan we have 3G phones already for about a year, since NTT Docomo introduced the FOMA service. It is hardly very successful, expensive and unreliable but these phones contain more features than any GSM phone on the market.
In addition KDDI subsidiary AU has sported a 2.5G phone service called CDMA-One for about 3 years, which now has been moved to full 3G with CDMA2000.
It is easy to think of GSM and Nokia as THE WORLD STANDARDS, but please remember that Japan has a completely different phone system that is much more advanced in many areas (while lagging behind in others).
In a typical Japanese bold strategic move, NTT declared recently that from now on it will only invest into VoIP equipment. Obviously they do not see a future in normal telephone services. As there are glass fibre connections available basically to every household in Tokyo, it makes sense to have just one line going in.
A 100M Fibre connection with one IP costs around 15,000 Yen (~130USD per month).
It seems that most people get around 60 Mbit out of this connection, with the bottleneck being the router.
Recently we upgraded a Siemens RM-400 machine running reliant unix. It took a Siemens engineer 14 hours to do the job.
What's the main problem with Unix installation's decay? - bad administration. Many user directories with Admins first names long past, logs in strange locations, different configuration mentalities. Most Linux guys on slash dot say something like "I know where my stuff is" or "I can just check xxx and fix it". This is because perhaps they are the only people using those machines as admins.
Also people tend not to clean up behind themselves, they create temporary directories, download some stuff and it remains there for ever.
Unfortunately most distributions also come with a lot of bloat (red hat, suse, etc...) by default. Easy enough to clean up, but it takes time. If you compile packages yourself, surely they are easier to remove completely, this is the main advantage of using Linux. But sometimes rpms break some other packages...
In the final analysis I have to say that Linux is relatively easier to clean up than windows. However, cleaning up Linux can also be a long and tedious job (and expensive!).
Slashdot Mirror
I suppose what the article means is that there are 4 x 48U racks installed in the server room. It is fiction that each rack could actually loaded with 48 x 1U servers! Potential problems are: cooling, weight, air (fire hazard), power supply.
Most likely actual rack usage looks as follows:
- Rack with 5 Servers
- Rack for Patching and switches
- Rack for phone system / phone patches
- Rack for backup.
If they have remaining capacity, they could rent it out/sell to other community organisations.
Sound to me like the poster uses a virtual shared server, ie. The hosting company provides a virtual root environment, this would be an explanation why they can "root" the server, ie. just access the file system directly from the hypervisor. I am sure this kind of service does not include uptime guarantees, it operates most likely under "Best Effort" promise. Which means they do not need to guarantee either uptime or availability. Their monitoring system and logs do not detect any error, so they want to check the posters systems logs - a reasonable request as they are trying to help him. Monitoring is a difficult task - what to you measure, where is your sensor. Was the outage on your own internet connection rather than on the server? "Moments ago, there were three simultaneous outages while I was logged into the server working on some projects. " It sounds very suspicious, like the admin interface causes interruption in the VM. Something like this happened to my VMWare Servers totally unexplainable, but reproducably.
Its a silly argument that the article makes. E-Readers can display B&W pages, and gray shades. This is absolutely sufficient for displaying textbooks. There is a line of extremely bloated textbooks that takes liberties with layout and colour. I just wonder for what purpose colour is used in those books. In my opinion it does neither help understanding of diagrams and models, nor does it improve the information density. Perhaps a revisit of Tufte's rules for information design is in order for publishers?
Companies operate with constrained resources in order to generate a profit. While developing a product a company only experiences cost, no income. After the selling starts a company receives income that slowly covers development cost until it breaks even. After break even, additional sales generate profit. After sales start any maintenance work reduces profit.
Why do I talk about what everyone already knows? A company like microsoft must decide for themselves how to use their limited resources. They decide how serious a bug is by looking at the urgency, and seriousness vs reduced profits. They make a decision based on their own interest. By putting an exploit on ebay, the cost and seriousness of the bug is not decided merely by microsoft, but by the market.
If MS thinks its irrelevant, they need not do anything. If they think its serious, they will have the choice of either fixing it until the auction ends, or bidding to prevent disclosure. Third parties interested in security will also have a chance to bid, however they must realise that their purchase may soon loose value if a fix is provided.
Overall I think its a good idea - in the best case it encourages a faster fixing of issues.
I agree. Converted from NT4 Server to Samba. There are some imperfection with oplocks, but samba itself works reliable in an NT Domain or Workgroup - after all Mac OS/x also uses Samba, and the author said that it has no problem!
Suse often supplies RPMs that do not work properly are configured in a strange way. If you are an ordinary customer they will not tell you the little secret that makes it work.
Saying that though, it should still be possible to make Suse's samba work. I suppose a configuration wizard would help this kind of user.
But to be fair, they tested the office version, so you would expect the essentials business apps to work straight away without tinkering.
Well seems that the guy who forked gentoo with his "great" gentoo embedded idea does not have much of a business proposition after all. And why should he, with competition from QNS and existing Linux embedded projects.
First of all, if I remember correctly Gallileo was conceived a long time ago, way before recent tensions between the US and Europe. It addressed the problems with the availability and accuracy of the American GPS system, so that it can be used (mainly) in Europe for civilian purposes: Automatic Freight unloading and transhipment in Ports, Aircraft & Ship navigation, as well as many more commercial activities.
What is the age of GPS satelites, and what is their useful working life? Are they really still state of the art? I think with Gallileo, the world has the chance to obtain a newer and better technology.
No matter what your political opinion, we should all welcome the better system. Some People are happy to run Suse or RedHat with low performance, others prefer optimised Gentoo. When the system is ready, you can still continue to use GPS if you want to. Nobody is forced to switch.
They are handmade, with a three year warranty and about as good as it gets. In my company we have hundreds of them and they never fail (both LCD & CRT).
"Someone" I know, just downloaded Freelancer - a Microsoft Game - and is really happy with it. He was thinking of buying the boxed version anyway, to get the a nice manual, and additional features like a printed starmap etc...
He then discovered on the MS Website, that the boxed version actually does NOT contain a manual!
Of course, now there is no reason to buy it anymore.
My conclusion is that people download warez games because of two reasons:
- They cannot get it in the shop (my friend who lives in Japan)
- There is less value in the product, but at the same price(Privateer and Wingcommander Prophecy had both Manuals and consisted of several CDs!!)
I guess your bad experience gives you the right to be scared of two wheel vehicles. But does it mean just because something is more dangerous than, lets say walking, one shouldn't do it anymore?
Knowing about the risks, one should perhaps decide on strategies to avoid the main problems: Avoid the main roads, rush hour, dangerous corners, etc...
Here in Tokyo I started to commute to work. I avoid the major ateries leading into the city centre, instead use small roads through residential areas & small paths along rivers. In very bad places I cycle on the sidewalk. On certain small roads with two way traffic, I block cars behind me from overtaking in dangerous places by cycling in the middle of the road.
Drivers here are so bad that, you may just as likely be hit walking! But then I suppose in LA, in addition to the dangers of traffic, you could also get shot if you go through the wrong neighbourhood?
Just purchased a road type bicycle on Friday. Wow, have things changed. The last time I rode a half decent bike was 15 years ago. It is all high tech mechanics nowadays: Aluminium Alloy light weight frame, Gear Levers integrated into the breaks, High pressure tires. As a result the thing accelerateslike hell even bearing my unfit body. Adrenalin rushes to your head, while you zip in and out of the traffic, overtaking traffic jammed cars - you are the true king of the road! How can a mere home training bike compete?
How about giving M$ a taste of their own medicine? NT users have to pay per server and per license fee, so why not charge charge M$ per square meter and per pedestrian that walked by at the time?
There is no point crying about all the tiny features that don't work in Suse. First of all, each distribution has a different way of doing things, which is nice as it shows that we have freedom of choice.
So when reviewing a distribution you need to look at it objectively from the value it gives you. People who have been using yast or yast 2 before are very effective in using it and working around the querks, just like many Linux user are beter with the bash or bourne shells than Windows users.
Secondly, most people are not interested in hearing the same old problem stories, after the authors keeps throwing all sorts of weird and exotic hardware-software combinations at the distro. For my part I am interested how well it works on normal PCs, eg. Dell Optiplex, Poweredge, Compaq Proliant servers, etc... I am sure in Germany (where Suse is located) there are many strange monitors that are 2 years old that are not auto-detected by RedHat et al! So saying that a particular monitor is not on the hardware list constitutes hardware compatibility issues is pretty daft(the same could be said about any linux system).
Also - graphical user interfaces are purely a matter of personal taste. If you don't like KDE, please use Gnome. It is no secret that Suse is KDE biased, so please do not blame them if you don't like the looks. It cannot be that difficult to download a theme or customize the desktop to fit your private preferences?
So at the end of the day, what value does Suse add?
- Grub substitutes Lilo, a necessary evolutionary step
- CUPS printing as default brings it up to the current state of the art
- XFS file system introduced already in 8.0 brings reliable journaling and ACLs for those people who do not like patching kernels and using ReiserFS or EXT3
- Apache, Samba, Bind and other standard apps precompiled with workable configurations
If you are a pro, you do not have to use these SUSE RPMs or yast, but instead you will compile these servers yourself anyway.
Overall Suse is not perfect, but neither are the other distros. Most problems are a matter of personal preference, but these problems also happen with other distros. The point is that Suse gets the job done, and once you take of your red hat and learn doing it the other way, you can actually get a nice result.
Well actually there is no such thing as a single 3G standard! As far as I know there are several competing standards mainly the European one and the American one (by Qualcom).
As a matter of fact here in Japan we have 3G phones already for about a year, since NTT Docomo introduced the FOMA service. It is hardly very successful, expensive and unreliable but these phones contain more features than any GSM phone on the market.
In addition KDDI subsidiary AU has sported a 2.5G phone service called CDMA-One for about 3 years, which now has been moved to full 3G with CDMA2000.
It is easy to think of GSM and Nokia as THE WORLD STANDARDS, but please remember that Japan has a completely different phone system that is much more advanced in many areas (while lagging behind in others).
In a typical Japanese bold strategic move, NTT declared recently that from now on it will only invest into VoIP equipment. Obviously they do not see a future in normal telephone services. As there are glass fibre connections available basically to every household in Tokyo, it makes sense to have just one line going in.
A 100M Fibre connection with one IP costs around 15,000 Yen (~130USD per month).
It seems that most people get around 60 Mbit out of this connection, with the bottleneck being the router.
Recently we upgraded a Siemens RM-400 machine running reliant unix. It took a Siemens engineer 14 hours to do the job.
...) by default. Easy enough to clean up, but it takes time. If you compile packages yourself, surely they are easier to remove completely, this is the main advantage of using Linux. But sometimes rpms break some other packages ...
What's the main problem with Unix installation's decay? - bad administration. Many user directories with Admins first names long past, logs in strange locations, different configuration mentalities. Most Linux guys on slash dot say something like "I know where my stuff is" or "I can just check xxx and fix it". This is because perhaps they are the only people using those machines as admins.
Also people tend not to clean up behind themselves, they create temporary directories, download some stuff and it remains there for ever.
Unfortunately most distributions also come with a lot of bloat (red hat, suse, etc
In the final analysis I have to say that Linux is relatively easier to clean up than windows. However, cleaning up Linux can also be a long and tedious job (and expensive!).