The great thing about them is there are so many to choose from. As someone once said.
The US TLD might be say.us with SLD of co , ac etc etc so you get for example novell.co.us . Then you realise that this is far too simple. Lets use the state abbreviation as well, so you now have a series of novell.co..us . Obviously the Internet is constrained by state boundaries (front line, DMZ or whatever you call them over there).
While we are at it why not rename.uk to.gb to follow the rest of the world's standards (GB is an ISO thing, UK isn't). I think there is another country that uses a non ISO TLD as well.
When it comes down to it, the.us domain should be relaunched but following the mould the rest of the world uses. That way the.com's.orgs etc could be used for truly international efforts (who would be the final arbitrator for that ?)
Finally, when playing the "guess that URL" game adding the twist of "guess the state and then remember its postal abbreviation then add on the nearest town's abbreviation and then pop the zip code on the end" probably doesn't endear the current.us domain to its users.
As to whether it is server or workstation "optimised", this is down to you. I run MD 7.0 both at work and home. The home setup is a wks with a small subset of servers running for the other PC (Win 98) and is very good - lots of toys etc etc.
The work box does a goodly amount of file serving. I use Samba to dole out antivirus updates (Sophos) to 3,500 NT PCs - this it does in an hour or so at 2-10Mb per client. This doesn't raise the run averages over 1 (it is a dual PIII 500 + 2Gb RAM). It does 802.1Q with a kernel patch to seven VLANs. There are normally up to 5 remote X sessions running and quite a few other things going on as well (eg NT server in a VMWare box). Up time is 80 odd days (I rebuilt the kernel to add VLAN tagging)
The point of the above diatribe is that I believe I have tested both sides of the coin and not found it wanting. Then again, if you spend long enough with any distro. or hand craft your own set up you can get all the bits together you need. However I have spent some time with other distros (admittedly older ones) and not found them to be so complete in terms of sensible customisations already applied out of the box. There are, of course still a few rough edges but these are surprisingly minimal.
Another point I would like to make is that none of the bleeding edge stuff in this distro have caused me problems (what does a kernel oops look like anyway ?) The only apps that have dumped core on me are stuff that I have applied and KDevelop (which I had given a hard time)
Finally, why not trundle over to http://lwn.net - they have links to virtually all (sensible) distros and possibly links to some reviews as well as on goig developments.
Personally speaking for a load of this magnitude SCSI is the only solution.
Don't even think of software RAID.
For some background on SCSI itself try http://www.scsifaq.org
There are many types of RAID 0-5 are the "standard" but there are several new ones eg level 10 which attempts to address throughput issues. Your actual space requirements don't seem outrageous so level 5 would be reasonably cost effective.
Another thing you will probably want is hot swapping. Once you've had a box tell you a drive is dead, you've removed it and popped a new one in without taking the box down, you will not want anything else.
On the IDE vs SCSI debate, whilst IDE is fast it seems to me that under continuous load SCSI gives better throughput.
As others have pointed out - a 'designed' server, rather than a "roll your own" box would make sense. Compaq Proliants make excelent Linux machines. The SMART arrays are very good and support RAID to level 5. You can fit a lot of disks in the drive cages as well. They are a little pricey but of a good quality and reliability. We have rather a lot of them running NetWare. I get to use the older kit to run my funny Open Source stuff...
A suggestion might be: Proliant 1600, 2 x 600Mhz processors, SMART 3200 with 64Mb cache, 5 drive slots - 81 Gb available after RAID 5 on 18Gb 1" drives (that's Ultra-2 SCSI) supports upto 1Gb RAM (has 128 by default). There is also an on-board SCSI interface for CDROM etc. This comes in at about GBP 9,000
The following worked for me when I first had symbol problems (RedHat and Mandrake):
Using 2.2.12 as an example:
If you are using the distro supplied source, edit the Makefile and make sure that the version numbers match the kernel release. Mandrake eg put MDx as the extra version number (or similar). This can cause you confusion - remove it so:
VERSION = 2
PATCHLEVEL = 2
SUBLEVEL = 12
EXTRA VERSION =
#make menuconfig
etc etc virtually everything as a module
#make dep
#make clean
#make bzImage
#make modules
(now mv/lib/modules/2.2.12
/lib/modules/2.2.12-old, do this whilst mods compiling)
Make sure that/boot/vmlinuz-old has an option just in case, to get you in if something horrible happens
After a reboot depmod should be run for you. uname -a should show you 2.2.12 as the kernel version, the modules should have installed as/lib/modules/2.2.12 and everything should be fine. RH and Mandy both have Linuxconf and various tools for managing the kernels but frankly I don't trust them and besides I don't like having fancy distro specific stuff in my kernels.
Also if you patch your own or download new ones they will work with the above recipe.
Slashdot Mirror
The great thing about them is there are so many to choose from. As someone once said.
.us with SLD of co , ac etc etc so you get for example novell.co.us . Then you realise that this is far too simple. Lets use the state abbreviation as well, so you now have a series of novell.co..us . Obviously the Internet is constrained by state boundaries (front line, DMZ or whatever you call them over there).
.uk to .gb to follow the rest of the world's standards (GB is an ISO thing, UK isn't). I think there is another country that uses a non ISO TLD as well.
.us domain should be relaunched but following the mould the rest of the world uses. That way the .com's .orgs etc could be used for truly international efforts (who would be the final arbitrator for that ?)
.us domain to its users.
The US TLD might be say
While we are at it why not rename
When it comes down to it, the
Finally, when playing the "guess that URL" game adding the twist of "guess the state and then remember its postal abbreviation then add on the nearest town's abbreviation and then pop the zip code on the end" probably doesn't endear the current
As to whether it is server or workstation "optimised", this is down to you. I run MD 7.0 both at work and home. The home setup is a wks with a small subset of servers running for the other PC (Win 98) and is very good - lots of toys etc etc.
The work box does a goodly amount of file serving. I use Samba to dole out antivirus updates (Sophos) to 3,500 NT PCs - this it does in an hour or so at 2-10Mb per client. This doesn't raise the run averages over 1 (it is a dual PIII 500 + 2Gb RAM). It does 802.1Q with a kernel patch to seven VLANs. There are normally up to 5 remote X sessions running and quite a few other things going on as well (eg NT server in a VMWare box). Up time is 80 odd days (I rebuilt the kernel to add VLAN tagging)
The point of the above diatribe is that I believe I have tested both sides of the coin and not found it wanting. Then again, if you spend long enough with any distro. or hand craft your own set up you can get all the bits together you need. However I have spent some time with other distros (admittedly older ones) and not found them to be so complete in terms of sensible customisations already applied out of the box.
There are, of course still a few rough edges but these are surprisingly minimal.
Another point I would like to make is that none of the bleeding edge stuff in this distro have caused me problems (what does a kernel oops look like anyway ?) The only apps that have dumped core on me are stuff that I have applied and KDevelop (which I had given a hard time)
Finally, why not trundle over to http://lwn.net - they have links to virtually all (sensible) distros and possibly links to some reviews as well as on goig developments.
Personally speaking for a load of this magnitude SCSI is the only solution.
...
Don't even think of software RAID.
For some background on SCSI itself try http://www.scsifaq.org
There are many types of RAID 0-5 are the "standard" but there are several new ones eg level 10 which attempts to address throughput issues. Your actual space requirements don't seem outrageous so level 5 would be reasonably cost effective.
Another thing you will probably want is hot swapping. Once you've had a box tell you a drive is dead, you've removed it and popped a new one in without taking the box down, you will not want anything else.
On the IDE vs SCSI debate, whilst IDE is fast it seems to me that under continuous load SCSI gives better throughput.
As others have pointed out - a 'designed' server, rather than a "roll your own" box would make sense. Compaq Proliants make excelent Linux machines. The SMART arrays are very good and support RAID to level 5. You can fit a lot of disks in the drive cages as well. They are a little pricey but of a good quality and reliability. We have rather a lot of them running NetWare. I get to use the older kit to run my funny Open Source stuff
A suggestion might be:
Proliant 1600, 2 x 600Mhz processors, SMART 3200 with 64Mb cache, 5 drive slots - 81 Gb available after RAID 5 on 18Gb 1" drives (that's Ultra-2 SCSI) supports upto 1Gb RAM (has 128 by default). There is also an on-board SCSI interface for CDROM etc. This comes in at about GBP 9,000
The following worked for me when I first had symbol problems (RedHat and Mandrake):
Using 2.2.12 as an example:
If you are using the distro supplied source, edit the Makefile and make sure that the version numbers match the kernel release. Mandrake eg put MDx as the extra version number (or similar). This can cause you confusion - remove it so:VERSION = 2
PATCHLEVEL = 2
SUBLEVEL = 12
EXTRA VERSION =
#make menuconfig
etc etc virtually everything as a module
#make dep
#make clean
#make bzImage
#make modules
(now mv /lib/modules/2.2.12
/lib/modules/2.2.12-old, do this whilst mods compiling)
#make modules_install
#cp /boot/vmlinuz vmlinuz-old
#cp /usr/src/linux/arch/i386/boot/bzImage /boot/vmlinuz
#cp /usr/src/linux/System.map /boot
now update /etc/lilo.conf and run /sbin/lilo
Make sure that /boot/vmlinuz-old has an option just in case, to get you in if something horrible happens
After a reboot depmod should be run for you. uname -a should show you 2.2.12 as the kernel version, the modules should have installed as /lib/modules/2.2.12 and everything should be fine. RH and Mandy both have Linuxconf and various tools for managing the kernels but frankly I don't trust them and besides I don't like having fancy distro specific stuff in my kernels.
Also if you patch your own or download new ones they will work with the above recipe.