then I'd suggest starting a wiki. Personally, I like and use Dokuwiki and have set up several processes to pull this sort of information directly into a wiki page (Dokuwiki uses plain.txt files) and it works great.
I've used LVM on my home server for a year or two now without any issues whatsover. I've been using it about as long on my development server at work as well...same story. Rock solid, no problems.
YMMV, however, as neither one of these boxes is heavily loaded and I've never required the functionality on any of my production servers..
If you're just trying to get sound (alerts, alarms , whatnot...) why not just buy some cheap speakers? You can pick up something that'll put out sound for under $10 per set. If you're looking for higher end sound (music, et all) then you're probably better off dropping the $5C.
We've been a Dell shop for over three years, but the 1650s ran us off. Various "mysterious" hardware problems (blue screens mostly pointing at the PERC or a hd failure) coupled with some really icky problems when we opened up the boxes: slot covers laying on the motherboard, fans not plugged in, loose screws, etc...
I think Dell's QA must have all been temporarily insane when the 1650s started rolling out.
All of our other Dell machines continue to work flawlessly day after day. Too bad really, we all thought the little blue light was wicked slick.
I'm guessing that you haven't actually implemented AD yet. As a linux admin in a very MS-biased shop, our AD implementation has single-handedly won over at least two previous MS-bigots to running linux machines at their desktops.
Multi-Master Replication: Replication leaves you with differing serial numbers on different servers. Updates to slave servers (you're not running primary query servers on your DCs are you?) still require you to configure notifies, or your updates will take hours (or whatever you have replication set to)to propagate to the other servers. Linux could replicate (punny =) this with LDAP, but why? DNS is perfectly capable of distributing that information all by it's lonesome. And it does it far faster than AD.
Win2k file/print/AD structure: I'm not sure what functionality is impressing you, but DNS & DHCP are an easy one. Printing is another no-brainer. Samba is far easier to configure than NT/2K ever dreamed of being and a whole mot more predictable. I get the impression that you make entirely too many changes to your production environment. Us linux bigots like to set stuff up and then forget it about for a couple of years.
DNS: Already kind of covered this, but my linux servers act as the primary query boxes and serve over a million internal hits a day (including all those RR and SRV records you seem so keen on). Add another million, each, every week or so for the external servers. Occassionaly, bind will need updating. Guess what, I don't have to reboot the whole box to do it. up2date bind (running RH), service named restart, and viola...downtime: approximately 5 seconds. Doing this on a win2k box means rebooting, and not just for the nameserver, for practically every friggin update. Dynamic updates? Also a no-brainer, if not an exceedingly good idea.
Printing: How flexible do you need your printing to be? Set up printer, point clients to printer, and print. How often do you have people moving around needing different printers? Do your drivers change frequently?
Now, the really pro-linux part. I currently have 7 servers ranging from a Dell PE2300 (dev server), 2 1550s (uptimes are currently over a year), 2 clustered 2650s (brand new, only up for 45 days or so), and a couple of workstations. My production boxes are sacrosanct as any outage requires signatures in triplicate from all the dept heads and a scheduled outage, usually at 2AM. I don't like being up at 2AM, so my boxes don't go down. I wanna see you update a Win2k box without rebooting it at least a third of the time. Our NT oracle clusters, while up for a year, had updates that really should have been done for almost that entire year.
My suggestion would be to hang out with some guys from your local LUG. Get a fresh perspective, as the MS brainwashing has obviously taken hold. Look outside the box a little bit.
I personally had a hot-cold RH7.3 cluster running on Dell 1650s for over two months before we trashed those boxes and sent them back to be exchanged for 2650s. Of the 17 1650s we bought, 11 had some nasty hardware problems, all the while Dell was telling us how none of their other customers were having problems with the line. The 1650 line is part of the reason we are now buying Compaq servers.
My 2650s have currently been up for about 45 days without a single problem. My 1550s have uptimes of over a year.
Hook a linux box up to the serial port and use minicom. Falling a bit more off-topic, we have some multi-port serial cards that the VMS guy has hooked up in a Windows pc and attached to all of his console ports. He can then remote into the one pc from anywhere, and have all of his console ports available. I need to find out if there's linux support for these cards, because I've become quite addicted to minicom for running my console sessions.
To explain a bit more fully: 1. Slow - our vpn is hosted at a remote office which means my users have the added latency of traversing the private circuits between the two offices to access resources in my datacenters. This leads to user complaints about speed, even though utilization on these lines is usually only 5-10%. 2. PITA to setup - I've seen the 13 page (printout, not terminal pages) configuration for the PIX at our other office. While that's not all VPN, a considerable amount is the nat/vpn setup. This may very well be excessive, but I'm glad it's not my PIX to manage. 3. flakey - I have users that constantly complain about lost connections, poor performance, and instability. This may be the configuration, as I do not manage that PIX. My only vpn tunnel is to a vendor via a 515 dedicated for the connection, and has been stable for about a year now with no intervention whatsoever. However, that connection uses a very small nat pool through a private network to a high-availability connection on the vendor's side and is only accessible to a very small number of servers on our side.
Now the preferred solution in my datacenter is ssh/radmin because: 1. Performance - I am able to access anything I require either directly from a shell on the gateway or via a tunnelled remote desktop. I have users that do spreadsheets from home and we receive very few complaints about screen refresh problem (which we had with vnc) or perceived "slowness". 2. ease of use - We've compiled a package of putty (freely available) and Remote Admin ($700 site license). It takes about 3 minutes to setup, including adding the user account to the gateway which authenticates off the domain, i.e., no user management on my side other than creating the initial account (I could automate it, but prefer the extra step). 3. stability - I have rock solid connections with good performance and the same choice of using remote software that I would at work (Win TS or RAdmin) all of which fits on a tiny flash disk or a couple of floppies. Of the 40 people using ssh/radmin, I receive fewer complaints than the 10 people that use the vpn. Most of my users only require telnet access to our alphas, so I created a limited shell which allows them telnet access from the inside of the gateway (on our private network) instead of setting up Reflections to use the vpn. I have created a limited shell which only allows certain commands from their login. The interface is easy to use and completely reliable from anywhere with any ssh client.
Additionally, I have some real concerns regarding the wide open connection that a vpn provides to our internal network. For this reason, we only allow vpn connections from company-owned and -administered machines. This allows us to guarantee that the remote machine is running up-to-date virus scanners and are not running unacceptable software. In the spirit of "only the access that the user needs", ssh/remote desktop works well, provides the users with what they need, and requires really minimal management and troubleshooting. The only real problem we've run into is trying to run this combo with WinXP Home, as it appears that M$ has broken something internally. WinXP Pro works fine, but Home simply refuses to run the remote software over the tunnel (heresay from workstation, I have no access to WinXP Home to troubleshoot).
Now, all that said, I am a huge fan of Cisco, having just dropped better than a quarter mil on new equipment for our new facility. Part of the new configuration will be a PIX 515EUR/FO bundle for testing site-site vpn to replace our costly long-haul private lines between offices (along with a 6513 and a couple of 7204s, I'm pretty psyced =).
We use two solutions depending on the client side hardware. On company-owned hardware (laptops mostly), they are allowed to use cisco VPN. Since the VPN is 1>slow, 2> a pita to set up, and 3>flakey we require an SSH/Remote Administrator combo on user-owned hardware. SSH to a gateway server handles most of the mainframe needs, and allows us to eliminate telnet connections directly from the outside while tunneling Tunneling Radmin allows them secured access to their desktops. Remote Admin is much faster than VNC (although not as fast as Terminal Server), and can be configured for NT authentication
The bill refers to "a technological measure", not encryption (however, I have not read the entire text of the bill). Since it's inception, one of the complaints about the DMCA is it's over-generalisation, why can't we leverage that for us? Why couldn't "a technological measure" refer to SpamAssasin? And taking it a little farther, copyright is: the exclusive legal right to reproduce, publish, and sell the matter and form (as of a literary, musical, or artistic work) -from Webster's Online. And lastly, my email address is not necessarily a fact, since I do run my own server and have the ability to change it whenever I like. Since I do run my own mail server, how much "creative license" is required before my email address becomes a work of art?
All I'm asking for is a little creative interpretation to generate some precedents that are favorable for ME.
Do you really think that the XBox was just to play games? Comeon, between XP Media Center and a $300 price tag, the XBox was aiming to be "the" box in your living room for both gaming and digital media, especially for homes without a PC already. IIRC, there were originally rumors of an add-on to the XBox for precisely that, although I haven't seen anything about that piece for a while. This would have pushed the MS DRM agenda nicely down the path. Unfortunately, their loss leader doesn't appear to be panning out quite as they'd like. A quick googling yeilds several results: here's a good one here's another and another
So, if I were to have some sort of genetic abnormality, such as a congenital heart defect or sickle-cell anemia, does that mean I would be "something other than human"? Does that mean I would no longer enjoy my "inalienable human rights" or have to pay taxes (since I'm now the equivalent of a dog)?
"The Beverage Network" merits it's own website complete with a news and reviews sections...I am, however, completely dissappointed in the lack of a Mountain Dew review. I mean, wth?
If you're comparing digital cash to online credit card transactions, then the difference is that digital cash is anonymous and irrepudiable.
If you're under the impression that ANY transaction you enact using your credit card is "anonymous", you are entirely mistaken. Having had to deal with the online credit card authentication process, I can assure you that the credit card company has a VERY good idea of who is purchasing what.
Common sense would force mw to agree with you. However, under the DMCA, it doesn't quite work that way. You are not allowed to circumvent copy protection schemes. This, IIRC, includes hardware.
Hence, "your" hardware really isn't "your" hardware in the sense that you can do whatever you like with it. Which begs the question, is taking an Xbox out in the backyard and smashing it with a sledgehammer considered circumventing their copy protection mechanisms?
I've been using OpenOffice for several months in a very MS-biased shop. Running on a 1.7P4 w/512, both office applications start blank in under 3 seconds. Specifically, I have a document where I track the email throughput on my linux gateways. Since others look at this document, I have always saved in MS format by default and it takes about 22 seconds to load. For s&g, I saved a copy in OO (sxc) and it takes 9 seconds to load. Opening with Excel in native format takes 4 seconds. So, in premilinary testing, yes, OO does take a bit longer to load, even in native format. The question then becomes, is it $400+ faster? Deciding the answer is a task left to the reader. Also, the xls document is 605K, while the same doc saved in OO is 68K. Zipping these docs yields sizes of 50K and 64K respectively. So who's format is more efficient?
I have been moderately active in the SCA (http://www.sca.org) in the past and game online ALOT. I feel that qualifies me for a response. When one first starts in the SCA, you are encouraged to pick a persona resembling someone that you would have liked to have been. This is similar to other reenactment/out-of-period groups. It can even be extended to online gaming and is even much easier in that venue. Some people spend HOURS, every day, in an online forum be it chat rooms, a MMORPG, or fragging bad guys in a shooter. It's hard to invest that much time in something and not invest some of yourself into it as well. We don't choose who we're born as, but we can certainly choose who we are.
Slashdot Mirror
The more they try to force me to do something that is not in my best interest, the more they marginalize themselves...
Wouldn't this also mean that the raven that I lost last night would be a deductible capital loss?
then I'd suggest starting a wiki. Personally, I like and use Dokuwiki and have set up several processes to pull this sort of information directly into a wiki page (Dokuwiki uses plain .txt files) and it works great.
I don't use alot of reiser fs's, but what do you mean by using reiser -instead- of LVM?
I've used LVM on my home server for a year or two now without any issues whatsover. I've been using it about as long on my development server at work as well...same story. Rock solid, no problems.
YMMV, however, as neither one of these boxes is heavily loaded and I've never required the functionality on any of my production servers..
perhaps you should try rm -rf /mnt/Windows?
If you're just trying to get sound (alerts, alarms , whatnot...) why not just buy some cheap speakers? You can pick up something that'll put out sound for under $10 per set.
If you're looking for higher end sound (music, et all) then you're probably better off dropping the $5C.
We've been a Dell shop for over three years, but the 1650s ran us off. Various "mysterious" hardware problems (blue screens mostly pointing at the PERC or a hd failure) coupled with some really icky problems when we opened up the boxes: slot covers laying on the motherboard, fans not plugged in, loose screws, etc...
I think Dell's QA must have all been temporarily insane when the 1650s started rolling out.
All of our other Dell machines continue to work flawlessly day after day. Too bad really, we all thought the little blue light was wicked slick.
I'm guessing that you haven't actually implemented AD yet. As a linux admin in a very MS-biased shop, our AD implementation has single-handedly won over at least two previous MS-bigots to running linux machines at their desktops.
Multi-Master Replication:
Replication leaves you with differing serial numbers on different servers. Updates to slave servers (you're not running primary query servers on your DCs are you?) still require you to configure notifies, or your updates will take hours (or whatever you have replication set to)to propagate to the other servers. Linux could replicate (punny =) this with LDAP, but why? DNS is perfectly capable of distributing that information all by it's lonesome. And it does it far faster than AD.
Win2k file/print/AD structure:
I'm not sure what functionality is impressing you, but DNS & DHCP are an easy one. Printing is another no-brainer. Samba is far easier to configure than NT/2K ever dreamed of being and a whole mot more predictable. I get the impression that you make entirely too many changes to your production environment. Us linux bigots like to set stuff up and then forget it about for a couple of years.
DNS:
Already kind of covered this, but my linux servers act as the primary query boxes and serve over a million internal hits a day (including all those RR and SRV records you seem so keen on). Add another million, each, every week or so for the external servers. Occassionaly, bind will need updating. Guess what, I don't have to reboot the whole box to do it. up2date bind (running RH), service named restart, and viola...downtime: approximately 5 seconds. Doing this on a win2k box means rebooting, and not just for the nameserver, for practically every friggin update. Dynamic updates? Also a no-brainer, if not an exceedingly good idea.
Printing:
How flexible do you need your printing to be? Set up printer, point clients to printer, and print. How often do you have people moving around needing different printers? Do your drivers change frequently?
Now, the really pro-linux part. I currently have 7 servers ranging from a Dell PE2300 (dev server), 2 1550s (uptimes are currently over a year), 2 clustered 2650s (brand new, only up for 45 days or so), and a couple of workstations. My production boxes are sacrosanct as any outage requires signatures in triplicate from all the dept heads and a scheduled outage, usually at 2AM. I don't like being up at 2AM, so my boxes don't go down. I wanna see you update a Win2k box without rebooting it at least a third of the time. Our NT oracle clusters, while up for a year, had updates that really should have been done for almost that entire year.
My suggestion would be to hang out with some guys from your local LUG. Get a fresh perspective, as the MS brainwashing has obviously taken hold. Look outside the box a little bit.
I personally had a hot-cold RH7.3 cluster running on Dell 1650s for over two months before we trashed those boxes and sent them back to be exchanged for 2650s. Of the 17 1650s we bought, 11 had some nasty hardware problems, all the while Dell was telling us how none of their other customers were having problems with the line. The 1650 line is part of the reason we are now buying Compaq servers.
My 2650s have currently been up for about 45 days without a single problem. My 1550s have uptimes of over a year.
Hook a linux box up to the serial port and use minicom. Falling a bit more off-topic, we have some multi-port serial cards that the VMS guy has hooked up in a Windows pc and attached to all of his console ports. He can then remote into the one pc from anywhere, and have all of his console ports available. I need to find out if there's linux support for these cards, because I've become quite addicted to minicom for running my console sessions.
To explain a bit more fully:
1. Slow - our vpn is hosted at a remote office which means my users have the added latency of traversing the private circuits between the two offices to access resources in my datacenters. This leads to user complaints about speed, even though utilization on these lines is usually only 5-10%.
2. PITA to setup - I've seen the 13 page (printout, not terminal pages) configuration for the PIX at our other office. While that's not all VPN, a considerable amount is the nat/vpn setup. This may very well be excessive, but I'm glad it's not my PIX to manage.
3. flakey - I have users that constantly complain about lost connections, poor performance, and instability. This may be the configuration, as I do not manage that PIX.
My only vpn tunnel is to a vendor via a 515 dedicated for the connection, and has been stable for about a year now with no intervention whatsoever. However, that connection uses a very small nat pool through a private network to a high-availability connection on the vendor's side and is only accessible to a very small number of servers on our side.
Now the preferred solution in my datacenter is ssh/radmin because:
1. Performance - I am able to access anything I require either directly from a shell on the gateway or via a tunnelled remote desktop. I have users that do spreadsheets from home and we receive very few complaints about screen refresh problem (which we had with vnc) or perceived "slowness".
2. ease of use - We've compiled a package of putty (freely available) and Remote Admin ($700 site license). It takes about 3 minutes to setup, including adding the user account to the gateway which authenticates off the domain, i.e., no user management on my side other than creating the initial account (I could automate it, but prefer the extra step).
3. stability - I have rock solid connections with good performance and the same choice of using remote software that I would at work (Win TS or RAdmin) all of which fits on a tiny flash disk or a couple of floppies. Of the 40 people using ssh/radmin, I receive fewer complaints than the 10 people that use the vpn. Most of my users only require telnet access to our alphas, so I created a limited shell which allows them telnet access from the inside of the gateway (on our private network) instead of setting up Reflections to use the vpn. I have created a limited shell which only allows certain commands from their login. The interface is easy to use and completely reliable from anywhere with any ssh client.
Additionally, I have some real concerns regarding the wide open connection that a vpn provides to our internal network. For this reason, we only allow vpn connections from company-owned and -administered machines. This allows us to guarantee that the remote machine is running up-to-date virus scanners and are not running unacceptable software. In the spirit of "only the access that the user needs", ssh/remote desktop works well, provides the users with what they need, and requires really minimal management and troubleshooting. The only real problem we've run into is trying to run this combo with WinXP Home, as it appears that M$ has broken something internally. WinXP Pro works fine, but Home simply refuses to run the remote software over the tunnel (heresay from workstation, I have no access to WinXP Home to troubleshoot).
Now, all that said, I am a huge fan of Cisco, having just dropped better than a quarter mil on new equipment for our new facility. Part of the new configuration will be a PIX 515EUR/FO bundle for testing site-site vpn to replace our costly long-haul private lines between offices (along with a 6513 and a couple of 7204s, I'm pretty psyced =).
We use two solutions depending on the client side hardware. On company-owned hardware (laptops mostly), they are allowed to use cisco VPN. Since the VPN is 1>slow, 2> a pita to set up, and 3>flakey we require an SSH/Remote Administrator combo on user-owned hardware. SSH to a gateway server handles most of the mainframe needs, and allows us to eliminate telnet connections directly from the outside while tunneling Tunneling Radmin allows them secured access to their desktops. Remote Admin is much faster than VNC (although not as fast as Terminal Server), and can be configured for NT authentication
All I'm asking for is a little creative interpretation to generate some precedents that are favorable for ME.
Do you really think that the XBox was just to play games? Comeon, between XP Media Center and a $300 price tag, the XBox was aiming to be "the" box in your living room for both gaming and digital media, especially for homes without a PC already. IIRC, there were originally rumors of an add-on to the XBox for precisely that, although I haven't seen anything about that piece for a while. This would have pushed the MS DRM agenda nicely down the path. Unfortunately, their loss leader doesn't appear to be panning out quite as they'd like.
A quick googling yeilds several results:
here's a good one
here's another
and another
1. deprive yourself of food, air, and water
2. ?????
3. PROFIT
how in the hell did you get mod'd to 2?
So, if I were to have some sort of genetic abnormality, such as a congenital heart defect or sickle-cell anemia, does that mean I would be "something other than human"? Does that mean I would no longer enjoy my "inalienable human rights" or have to pay taxes (since I'm now the equivalent of a dog)?
"The Beverage Network" merits it's own website complete with a news and reviews sections...I am, however, completely dissappointed in the lack of a Mountain Dew review. I mean, wth?
Since you're already familiar with RH, why not take a look at AS 2.1? IIRC, it's based on Mission Critical Linux's Kimberlite config.
Microsoft has knowingly violated our trust before...
If you're comparing digital cash to online credit card transactions, then the difference is that digital cash is anonymous and irrepudiable.
If you're under the impression that ANY transaction you enact using your credit card is "anonymous", you are entirely mistaken. Having had to deal with the online credit card authentication process, I can assure you that the credit card company has a VERY good idea of who is purchasing what.
Common sense would force mw to agree with you. However, under the DMCA, it doesn't quite work that way. You are not allowed to circumvent copy protection schemes. This, IIRC, includes hardware.
Hence, "your" hardware really isn't "your" hardware in the sense that you can do whatever you like with it. Which begs the question, is taking an Xbox out in the backyard and smashing it with a sledgehammer considered circumventing their copy protection mechanisms?
I've been using OpenOffice for several months in a very MS-biased shop. Running on a 1.7P4 w/512, both office applications start blank in under 3 seconds. Specifically, I have a document where I track the email throughput on my linux gateways. Since others look at this document, I have always saved in MS format by default and it takes about 22 seconds to load. For s&g, I saved a copy in OO (sxc) and it takes 9 seconds to load. Opening with Excel in native format takes 4 seconds. So, in premilinary testing, yes, OO does take a bit longer to load, even in native format. The question then becomes, is it $400+ faster? Deciding the answer is a task left to the reader.
Also, the xls document is 605K, while the same doc saved in OO is 68K. Zipping these docs yields sizes of 50K and 64K respectively. So who's format is more efficient?
I have been moderately active in the SCA (http://www.sca.org) in the past and game online ALOT. I feel that qualifies me for a response.
When one first starts in the SCA, you are encouraged to pick a persona resembling someone that you would have liked to have been. This is similar to other reenactment/out-of-period groups. It can even be extended to online gaming and is even much easier in that venue. Some people spend HOURS, every day, in an online forum be it chat rooms, a MMORPG, or fragging bad guys in a shooter. It's hard to invest that much time in something and not invest some of yourself into it as well.
We don't choose who we're born as, but we can certainly choose who we are.