The questions raised however, come from Perl and MySQL. Both are questionable in terms of scalability. Although I'm not qualified to comment on this, I belive that the general concensus is that MySQL is one of the least efficent databases today.
Well, you certainly proved your statement that you're not qualified to comment on this. Why should that stop you, though?
You do realize that Slashdot is built on MySQL/Perl (well, DB2 now for political reasons, but MySQL for a very long time), as is Movable Type and Typepad from Six Apart? I'm not sure how a network infrastructure can be "professional", or how LiveJournal's architecture doesn't qualify. Then again, I'm not sure how you would be qualified to just what is and is not "professional".
LiveJournal is more interesting than you think
on
LiveJournal Buyout Rumor
·
· Score: 4, Informative
Most geeks seem to react to hearing "LiveJournal" with something along the lines of "haha, livejournal sucks! it's just a bunch of 12-year-old girls complaining about their parents!" However, the service is quite interesting from a geek perspective:
They run a pretty huge web application (700-800 pageviews per second at peak, most of them database-backed), and Brad has written quite a bit about the challenges and solutions they've come up with. They've also written several very interesting open source infrastructure applications like memcached (used by Slashdot) and perlbal.
Thus, while the service may not be all that interesting, the tech behind it certainly is (at least to this geek).
Sun parts cost more. Everyone knows that. Where you win is when something breaks. If you put a Sun part in a Sun server, that part is covered under your support contract. If you put a third-party part in, it isn't. The difference between a $200 Crucial DIMM and a $800 Sun DIMM is not hearing "that isn't a supported part. yank it out and see if it solves the problem."
You're not paying extra for the part (like you said, most of it is re-badged from other manufacturers anyway). You're paying extra to have the new part covered under service.
Your example doesn't really hold. You're talking about suspending a process, modifying the binary, then un-suspending.
My instructions were for replacing the binary of a daemon process that is:
a) not suspendeed
b) most likely not swapped to disk, since you're actively talking to it via your ssh session.
There's no reason for the OS to ever need to read the binary from disk again. Even if it was swapped out it would read the image from swap, not the binary from disk. Again, like I said in my post, I've done this many times on Solaris, OpenBSD, and Linux without issue. Haven't tried SunOS since I don't have access to machines old enough to be running it.
Re:*** Help on upgrading a remote server?
on
OpenSSH Local Root Hole
·
· Score: 5, Informative
Known to work on Solaris, OpenBSD, and Linux. YMMV elsewhere, but it should work fine.
1) Use SSH to log into your server.
2) Install the new ssh version. Your old version is in memory, so replacing the binary won't have any adverse effect on your connection.
3) Run 'ps -ef | grep sshd' or 'ps auxw | grep sshd' (depending on your UNIX flavor)
4) find the sshd instance with a parent process ID of '1' -- this will be the actual daemon spawend by init. The other process will be the one spawned by sshd itself to handle your connection.
5) kill
6) the parent sshd process will terminate, but yours will stay running
7) start up the new sshd
8) from another workstation or window telnet to port 22 on your server and verify that the version number reflects the new version.
9) from another workstatino or window, ssh into your server to make sure you still have access.
10) close your original ssh session
I've used this exact process to upgrade many machines at remote locations. As long as you verify that the new sshd is running before you close your existing connection, you should have no problems.
One thing I personally like about John Carmack and id software is that they have a long history of (eventually) releasing their games as open source.
Take a look at the FTP site: Wolfenstein 3D, Doom, Quake, and Quake 2 are all available. You still have to have the map files and other game data from a "real" copy of the game in order to play it, but all of Carmack and Co's magic is up there for study.
In short, they have quite a history of "giving back to the community." Even for games (such as Q3) where the full source isn't released, id always releases SDK's (for lack of a better term) to allow anyone that wants to the chance to create add-ons, extensions, and "total conversions" -- new games based on the existing code.
"Or, we don't like you because your client does stuff that would hurt us finacially?"
What exactly do you expect a business (much less a public company) to do? It's their service, they have the right to decide who uses it. Personally, I hate what AOL's doing, but don't pretend that they're providing AIM out of the goodness of their hearts. They expect to make money off it (ads).
So the Ximian Installer is borked because you can't RTFM? Ximian has not yet released support for Mandrake 8.1, which you would have known if you had read their web site. Their OS and distribution support is clearly listed in several places on the site.
* Debian GNU/Linux 2.2 (Potato) on x86
* Linux Mandrake 7.0, 7.1, 7.2, 8.0
* LinuxPPC 2000
* Red Hat Linux 6.2, 7.0, 7.1, 7.2 on x86
* Solaris 7, 8 on UltraSPARC
* SuSE 6.4, 7.0, 7.1, 7.2 on x86
* Turbolinux 6.0
* Yellow Dog Linux 1.2, 2.0
Just curious as to why RedHat decided to ship old version of Perl and Python. Perl is 5.6.0 (5.6.1 has been out for quite a while), and Python is 1.5.2 (we're on the 2.2 betas).
Unfortunately, I can't, because Mandrake 8.1 isn't supported yet. Any idea when it might be? Seems like I recall Mandrake 8.0 support taking a while after the final release of the OS.
Cisco just announced the Pix 501, targeted at SOHO, but running the same PixOS as the "big iron" Pix firewalls. I'd be very surprised if it doesn't do everything you want.
This same subject comes up every time TiVo is mentioned on Slashdot.
Yes, they collect data on what you watch
Yes, they're upfront about this (it's in the manual. not their fault you didn't read it).
No, they don't tie the data back to you in any of the information they give out.
No, their terms of service (that contract you agreed to) says they can't change their minds and start selling your viewing data with identifiers attached.
Three words: Service Level Agreement. Any reputible ISP that sells you a T1 also sells you a SLA with it. If the line is down more than X minutes per month, you get $Y in compensation (usually off your bill).
Take a look at any DSL service agreement. I guarantee you'll see the words "best effort," meaning the ISP doesn't guarantee *any* level of service. Sure, they'll probably credit you if your service is down (and you call to complain) *but they don't have to!*. Think you're guaranteed 1.5mb/256mb because that's what the ISP advertises? *think again* Reading BellSouth's (the telco, not bs.net) technical specification for DSL local loops (house to DSLAM), they only guarantee 256k/sec as the point where they declare a loop good or bad. Anything over 256k/sec is "best effort"
In short, if you're a business and you rely on the Internet as part of your core, pay the extra to an established ISP and get a T1 with a decent SLA.
The company I work for currently puts an Ultra 5 in each office for DHCP & secondary DNS. We'll be looking hard at these as a way to save cash as we expand.
The large capacity Tivo only stores 30 hours of video (and this is at Basic, read almost unwatchable, quality).
My "hacked" Tivo, with a 15GB main drive (the 14 hour model) and a 75GB secondary drive can record over 90 hours at Basic and nearly 60 hours at Medium (which I actually use).
In short, you "hack" the Tivo to get more capacity than is available from the store. Sort of a reward for having the skill to do it.
>What's wrong with it? It's a devel version - it says as much in the story.
It also says "If you haven't upgraded yet, now is the time." - irresponsible, since it may lead folks to download and install a potentially VERY unstable release.
Actually, I'm reasonably sure that the "Net Force" series of books is just like the "Op Force" series: Clancy wrote none of them, only lending his name to the series for marketing purposes and perhaps acting as a consultant. I believe all these books are written by others. Clancy has posted to Usenet saying as much.
It's very commonplace to have multiple load-balanced web servers talking to a single NFS-mounted disk storage backend. It prevents problems with replicating content to all your web servers. At my company, we use a farm of 32 or so Sun web servers all sharing a NetApp filer for storage. All of your files and scripts are pulled from the same source. Plus, you don't have to invest in huge amounts of disk for each of your web servers. You just put your money into the NFS server. Also makes backups much easier. Cheers, Brian
If there were no banner ads, how many of your favorite web sites would no longer exist? I can think of one example: SLASHDOT. Do you really think Andover would have bought slashdot.org and poured money into running it if there was not a steady stream of ad revenue coming in? How about news.com, freshmeat.net, epinions, or many other content- or community-centric web sites? These companies have to have revenue to operate. Ad banners are in many cases the main or sole source of that revenue. Blocking ads at the browser level is irresponsible to the Web community.
Slashdot Mirror
Ask Sony how well having a ready-made mythology went with Star Wars Galaxies. Anything can be screwed up.
The questions raised however, come from Perl and MySQL. Both are questionable in terms of scalability. Although I'm not qualified to comment on this, I belive that the general concensus is that MySQL is one of the least efficent databases today.
Well, you certainly proved your statement that you're not qualified to comment on this. Why should that stop you, though?
You do realize that Slashdot is built on MySQL/Perl (well, DB2 now for political reasons, but MySQL for a very long time), as is Movable Type and Typepad from Six Apart? I'm not sure how a network infrastructure can be "professional", or how LiveJournal's architecture doesn't qualify. Then again, I'm not sure how you would be qualified to just what is and is not "professional".
Most geeks seem to react to hearing "LiveJournal" with something along the lines of "haha, livejournal sucks! it's just a bunch of 12-year-old girls complaining about their parents!" However, the service is quite interesting from a geek perspective: They run a pretty huge web application (700-800 pageviews per second at peak, most of them database-backed), and Brad has written quite a bit about the challenges and solutions they've come up with. They've also written several very interesting open source infrastructure applications like memcached (used by Slashdot) and perlbal. Thus, while the service may not be all that interesting, the tech behind it certainly is (at least to this geek).
Which will keep you safe exactly as long as it takes a judge to order you to give up your encryption key. Refuse and go to prison.
Actually, .com, .org, and .net servers aren't on the roots. They're on the gtld-server.net servers, which are completely separate.
There's a pretty old bug (really a feature request) open in Mozilla's bugzilla for adding NTLM authentication. One day soon, maybe!
Sun parts cost more. Everyone knows that. Where you win is when something breaks. If you put a Sun part in a Sun server, that part is covered under your support contract. If you put a third-party part in, it isn't. The difference between a $200 Crucial DIMM and a $800 Sun DIMM is not hearing "that isn't a supported part. yank it out and see if it solves the problem."
You're not paying extra for the part (like you said, most of it is re-badged from other manufacturers anyway). You're paying extra to have the new part covered under service.
Your example doesn't really hold. You're talking about suspending a process, modifying the binary, then un-suspending.
My instructions were for replacing the binary of a daemon process that is:
a) not suspendeed
b) most likely not swapped to disk, since you're actively talking to it via your ssh session.
There's no reason for the OS to ever need to read the binary from disk again. Even if it was swapped out it would read the image from swap, not the binary from disk. Again, like I said in my post, I've done this many times on Solaris, OpenBSD, and Linux without issue. Haven't tried SunOS since I don't have access to machines old enough to be running it.
Known to work on Solaris, OpenBSD, and Linux. YMMV elsewhere, but it should work fine.
1) Use SSH to log into your server.
2) Install the new ssh version. Your old version is in memory, so replacing the binary won't have any adverse effect on your connection.
3) Run 'ps -ef | grep sshd' or 'ps auxw | grep sshd' (depending on your UNIX flavor)
4) find the sshd instance with a parent process ID of '1' -- this will be the actual daemon spawend by init. The other process will be the one spawned by sshd itself to handle your connection.
5) kill
6) the parent sshd process will terminate, but yours will stay running
7) start up the new sshd
8) from another workstation or window telnet to port 22 on your server and verify that the version number reflects the new version.
9) from another workstatino or window, ssh into your server to make sure you still have access.
10) close your original ssh session
I've used this exact process to upgrade many machines at remote locations. As long as you verify that the new sshd is running before you close your existing connection, you should have no problems.
Take a look at the FTP site: Wolfenstein 3D, Doom, Quake, and Quake 2 are all available. You still have to have the map files and other game data from a "real" copy of the game in order to play it, but all of Carmack and Co's magic is up there for study.
In short, they have quite a history of "giving back to the community." Even for games (such as Q3) where the full source isn't released, id always releases SDK's (for lack of a better term) to allow anyone that wants to the chance to create add-ons, extensions, and "total conversions" -- new games based on the existing code.
Very, very, cool.
"Or, we don't like you because your client does stuff that would hurt us finacially?"
What exactly do you expect a business (much less a public company) to do? It's their service, they have the right to decide who uses it. Personally, I hate what AOL's doing, but don't pretend that they're providing AIM out of the goodness of their hearts. They expect to make money off it (ads).
So the Ximian Installer is borked because you can't RTFM? Ximian has not yet released support for Mandrake 8.1, which you would have known if you had read their web site. Their OS and distribution support is clearly listed in several places on the site.
* Debian GNU/Linux 2.2 (Potato) on x86
* Linux Mandrake 7.0, 7.1, 7.2, 8.0
* LinuxPPC 2000
* Red Hat Linux 6.2, 7.0, 7.1, 7.2 on x86
* Solaris 7, 8 on UltraSPARC
* SuSE 6.4, 7.0, 7.1, 7.2 on x86
* Turbolinux 6.0
* Yellow Dog Linux 1.2, 2.0
Just curious as to why RedHat decided to ship old version of Perl and Python. Perl is 5.6.0 (5.6.1 has been out for quite a while), and Python is 1.5.2 (we're on the 2.2 betas).
I'm sticking with Mandrake 8.1 for now.
Unfortunately, I can't, because Mandrake 8.1 isn't supported yet. Any idea when it might be? Seems like I recall Mandrake 8.0 support taking a while after the final release of the OS.
Cisco product information is here.
You already have no privacy. Get over it.
Three words: Service Level Agreement. Any reputible ISP that sells you a T1 also sells you a SLA with it. If the line is down more than X minutes per month, you get $Y in compensation (usually off your bill).
Take a look at any DSL service agreement. I guarantee you'll see the words "best effort," meaning the ISP doesn't guarantee *any* level of service. Sure, they'll probably credit you if your service is down (and you call to complain) *but they don't have to!*. Think you're guaranteed 1.5mb/256mb because that's what the ISP advertises? *think again* Reading BellSouth's (the telco, not bs.net) technical specification for DSL local loops (house to DSLAM), they only guarantee 256k/sec as the point where they declare a loop good or bad. Anything over 256k/sec is "best effort"
In short, if you're a business and you rely on the Internet as part of your core, pay the extra to an established ISP and get a T1 with a decent SLA.
The company I work for currently puts an Ultra 5 in each office for DHCP & secondary DNS. We'll be looking hard at these as a way to save cash as we expand.
My "hacked" Tivo, with a 15GB main drive (the 14 hour model) and a 75GB secondary drive can record over 90 hours at Basic and nearly 60 hours at Medium (which I actually use).
In short, you "hack" the Tivo to get more capacity than is available from the store. Sort of a reward for having the skill to do it.
>What's wrong with it? It's a devel version - it says as much in the story.
It also says "If you haven't upgraded yet, now is the time." - irresponsible, since it may lead folks to download and install a potentially VERY unstable release.
For an interesting take on "rewrite it from scratch" software development, check this out:
http://joel.editthispage.com/st ories/storyReader$47
Talks about Mozilla, Borland, and other projects that have had serious problems rewriting from scratch. Also,
http://joel.editthispage.com/2000/05/26 (scroll down)
A great quote from Lou Montulli of the original Netscape team.
Actually, I'm reasonably sure that the "Net Force" series of books is just like the "Op Force" series: Clancy wrote none of them, only lending his name to the series for marketing purposes and perhaps acting as a consultant. I believe all these books are written by others. Clancy has posted to Usenet saying as much.
They noted in the first article that Exodus does not allow cameras inside their facilities. Boo.
It's very commonplace to have multiple load-balanced web servers talking to a single NFS-mounted disk storage backend. It prevents problems with replicating content to all your web servers. At my company, we use a farm of 32 or so Sun web servers all sharing a NetApp filer for storage. All of your files and scripts are pulled from the same source. Plus, you don't have to invest in huge amounts of disk for each of your web servers. You just put your money into the NFS server. Also makes backups much easier. Cheers, Brian
If there were no banner ads, how many of your favorite web sites would no longer exist? I can think of one example: SLASHDOT. Do you really think Andover would have bought slashdot.org and poured money into running it if there was not a steady stream of ad revenue coming in? How about news.com, freshmeat.net, epinions, or many other content- or community-centric web sites? These companies have to have revenue to operate. Ad banners are in many cases the main or sole source of that revenue. Blocking ads at the browser level is irresponsible to the Web community.