I highly appreciate everyone's help in the caching name server bit. Building on that one:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network. Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg).
Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
William Warren hescominsoon@emmanuelcomputerconsulting.com wrote:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network.
This is probably best.
If you're using Samba, you should probably setup WINS with the same server as a WINS-to-DNS proxy as fallback.
Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg). Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
You can setup BIND 9 (and some 8.x versions?) to do Dynamic DNS (DDNS). Windows clients should attempt to register their names in the DDNS server by default.
I actually told Samba at setup to act as a Wins server. I have my Astaro firewall DHCP telling my clients that Enoch is the Wins server. That's all i did though. i did not setup or tweak anything else yet as i wanted to get DNS setup.
Bryan J. Smith wrote:
William Warren hescominsoon@emmanuelcomputerconsulting.com wrote:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network.
This is probably best.
If you're using Samba, you should probably setup WINS with the same server as a WINS-to-DNS proxy as fallback.
Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg). Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
You can setup BIND 9 (and some 8.x versions?) to do Dynamic DNS (DDNS). Windows clients should attempt to register their names in the DDNS server by default.
On Wed, 2005-08-10 at 13:15, William Warren wrote:
I highly appreciate everyone's help in the caching name server bit. Building on that one:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network. Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg).
I think the only difference is that you add the master entries in your named.conf file and appropriate zone files for the data for your internal zones. I'd recommend trying webmin for this because it will save the grunge work of building the reverse tables.
Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
You can make dhcp do dynamic updates, but for a small network the simple-minded way is to put the names and IP numbers in DNS, then put the names in the dhcpd.conf entries tied to mac addresses. That way if you want to change an address, you do it in the dns data and the dhcp will pick it up.
By the way, if you really don't want to know all this stuff, you might like the SME server from www.contribs.org. The next version is going to be based on Centos4 and there is an alpha release available now. There are things I don't like about the system but for small networks where one server does everything and you want 'fill-in-the-form' administration it is a good fit.
Les Mikesell lesmikesell@gmail.com wrote:
By the way, if you really don't want to know all this stuff, you might like the SME server from www.contribs.org. The next version is going to be based on Centos4 and there is an alpha release available now. There are things I
don't
like about the system but for small networks where one server does everything and you want 'fill-in-the-form' administration it is a good fit.
In other words, it's great for deploying at companies when you can't be around -- like most small consultants are for at small companies.
The last time I checked out work on SME, a number of Debian users were complaining that it should be Debian based. I have like Debian (Murdock's attitudes, used to be a maintainer too, etc...), but I thought it was rather self-defeating to move away from all the work that had already been done for the Red Hat Linux / RPM lineage. Especially since those people were not very into the existing codebase (or support).
I think it's great that they are going to base it on CentOS 4! And I'm glad to hear there is an Alpha test available.
On Wed, 2005-08-10 at 13:46, Bryan J. Smith wrote:
By the way, if you really don't want to know all this stuff, you might like the SME server from www.contribs.org. The next version is going to be based on Centos4 and there is an alpha release available now. There are things I
don't
like about the system but for small networks where one server does everything and you want 'fill-in-the-form' administration it is a good fit.
In other words, it's great for deploying at companies when you can't be around -- like most small consultants are for at small companies.
Yes, it installs as an 'appliance' that does about everything you'd want a single server to do. I've used it in remote offices where a local manager could take care of it himself.
The last time I checked out work on SME, a number of Debian users were complaining that it should be Debian based. I have like Debian (Murdock's attitudes, used to be a maintainer too, etc...), but I thought it was rather self-defeating to move away from all the work that had already been done for the Red Hat Linux / RPM lineage. Especially since those people were not very into the existing codebase (or support).
I never liked debian until Knoppix and Ubuntu turned the codebase into a very undebian-like product so those complaints never made sense to me. I have my own, though, which basically boil down to the system being completely configured by perl scripts and templates. If you want to make a change that the scripts don't handle (and I always did...), you not only had to understand the config files you were trying to avoid, but you also had to know perl and how all the related variables in the database are managed by the web forms. And up until now, it has been difficult to keep a system updated. Of course if you are a contractor selling support you don't want the users doing that kind of stuff themselves anyway so my complaints on their own development list were never well received.
I think it's great that they are going to base it on CentOS 4! And I'm glad to hear there is an Alpha test available.
It looks like they are going to make it use stock rpms at least to a certain extent - they still plan to use qmail and some of the other djb stuff, but at least a normal yum update will work. But, the ways they simplify administration combine some concepts in ways that are sometimes handy and sometimes just wrong - for example their DHCP server hands itself out as the DNS server and the DNS server is automatically configured as the primary for it's domain. In a larger network you are likely to already have a primary dns server for the domain but the SME clients won't see it. On the other hand, having a single form where you can enter a mac address along with the name/ip and have both the dhcpd.conf and dns zone file entry built is nice where it works.
actually I WANT to know this stuff. It took me weeks to figure out basic NTP. I could have gone the "easy" route but I am glad i didn't because now i can help others setup their own linux NTP servers.
I am not sure the Astaro's DHCP does dynamic updates?(could you explain that or is it literally as it seems?) i will ask them about that one.
So basically to start off i will have to enter the names and ip addresses manually. That's fine as the network itself is small. I am going to be adding servers and clients to it though
2 more clients running linux or windows for general usage 4 clients running mythtv in client mode 1 myth tv media server 2 notebooks(one will be to replace the gateway and one is for me).
This is not all at once..this is long range.
Les Mikesell wrote:
On Wed, 2005-08-10 at 13:15, William Warren wrote:
I highly appreciate everyone's help in the caching name server bit. Building on that one:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network. Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg).
I think the only difference is that you add the master entries in your named.conf file and appropriate zone files for the data for your internal zones. I'd recommend trying webmin for this because it will save the grunge work of building the reverse tables.
Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
You can make dhcp do dynamic updates, but for a small network the simple-minded way is to put the names and IP numbers in DNS, then put the names in the dhcpd.conf entries tied to mac addresses. That way if you want to change an address, you do it in the dns data and the dhcp will pick it up.
By the way, if you really don't want to know all this stuff, you might like the SME server from www.contribs.org. The next version is going to be based on Centos4 and there is an alpha release available now. There are things I don't like about the system but for small networks where one server does everything and you want 'fill-in-the-form' administration it is a good fit.
On Wed, 2005-08-10 at 13:48, William Warren wrote:
actually I WANT to know this stuff. It took me weeks to figure out basic NTP. I could have gone the "easy" route but I am glad i didn't because now i can help others setup their own linux NTP servers.
Errr... There isn't that much that is useful for most people to know about NTP other than how to point it at some servers. Running the GUI version of 'dateconfig' will do the right thing if you click the 'network time protocol' tab and fill in the form.
I am not sure the Astaro's DHCP does dynamic updates?(could you explain that or is it literally as it seems?) i will ask them about that one.
There is a protocol used to send dynamic updates to a DNS server. It was invented for people that don't know/care how names are assigned and just hope it all works. There is a security mechanism that can be used if you care a little bit - but setting it all up can be harder than assigning the names in the first place.
So basically to start off i will have to enter the names and ip addresses manually. That's fine as the network itself is small. I am going to be adding servers and clients to it though
With servers, you always have to assign the IP's. Clients can be dynamic and you can get by with generic names for their reverse lookups (dhcp-nn.nn.nn.nn, etc.) unless you want log files, packet sniffers, etc. to always have meaningful host names.
2 more clients running linux or windows for general usage 4 clients running mythtv in client mode 1 myth tv media server 2 notebooks(one will be to replace the gateway and one is for me).
This is not all at once..this is long range.
I'd type the addresses into webmin until you get a number you can't manage that way. After a thousand or so you might have trouble.
Folks. I thought this was a CentOS list not a Bind list... This list is getting out of hand lately.
Just my 2 cents,
--jesse
On Thursday 11 August 2005 01:52, Jesse wrote:
Folks. I thought this was a CentOS list not a Bind list... This list is getting out of hand lately.
You know, in a way I agree with this sentiment.
However, at the same time I also see that discussing the CentOS distribution of bind is not a big deal for this list.
As to the idea that, if people want to talk about BIND they should go to a BIND list, I would pose this question: are people to have to subscribe to a couple of thousand mailing lists, each of which talks about a specific component distributed with CentOS, and really flood their inboxes, or should we be patient with a couple of dozen component-specific threads in this list, which is likely to be friendlier to CentOS-related questions on that component? What is wrong with being patient, and courteous?
Yes, if a thread goes too far, then perhaps it and its participants should discuss on a more specific mailing list, or privately. But in my opinion at least this BIND thread hasn't gotten that far, and, for that matter, would probably not be on-topic for any of the BIND lists either.
But what is CentOS but a distribution of component packages? If we can't discuss issues with the components, what use is the list? Discussing the vagaries of the way bootable ISO's and yum repositories work? Discussing the very specific patches and mods made by Red Hat upstream of the CentOS build process (but that would be more appropriate to nahant-list or taroon-list, right?)? What IS on-topic for the list, then?
In my opinion, it is nice to have a relatively small, CentOS-friendly community with which to discuss component issues; many of the upstream packages' lists would be much less friendly (LKML anyone?).
And the fact of the matter is, this list is pretty low traffic.
} } However, at the same time I also see that discussing the CentOS } distribution } of bind is not a big deal for this list. } } As to the idea that, if people want to talk about BIND they should go to a } BIND list, I would pose this question: are people to have to subscribe to } a } couple of thousand mailing lists, each of which talks about a specific } component distributed with CentOS, and really flood their inboxes, or } should } we be patient with a couple of dozen component-specific threads in this } list, } which is likely to be friendlier to CentOS-related questions on that } component? What is wrong with being patient, and courteous? } } Yes, if a thread goes too far, then perhaps it and its participants should } discuss on a more specific mailing list, or privately. But in my opinion } at } least this BIND thread hasn't gotten that far, and, for that matter, would } probably not be on-topic for any of the BIND lists either. } } But what is CentOS but a distribution of component packages? If we can't } discuss issues with the components, what use is the list? Discussing the } vagaries of the way bootable ISO's and yum repositories work? Discussing } the } very specific patches and mods made by Red Hat upstream of the CentOS } build } process (but that would be more appropriate to nahant-list or taroon-list, } right?)? What IS on-topic for the list, then? } } -- } Lamar Owen
right on Lamar,
good thoughts and some key words there like patient and courteous.
what i notice is that some or much of the time people forget how _easy_ it is to be rude to others and how incredibly challenging and/or difficult it can be to be nice to other people...
i on the other hand have to remember that i need not treat every situation on this "list" or in "life" like a "very serious network down emergency"... sheesh... can anyone relate? ;->
i guess when we help others, we want to help so bad sometimes that we might become somewhat ineffective when we "experts" do not actually have our hands on the equipments with the _illness_ so to speak.
-- Robert Hanson Abba Communications http://www.abbacomm.net
On Thu, 2005-08-11 at 08:55, Lamar Owen wrote:
On Thursday 11 August 2005 01:52, Jesse wrote:
Folks. I thought this was a CentOS list not a Bind list... This list is getting out of hand lately.
You know, in a way I agree with this sentiment.
However, at the same time I also see that discussing the CentOS distribution of bind is not a big deal for this list.
There was a distribution-packaging part of this discussion that wasn't answered. If someone wants to bring it back on track they could explain what you get with an 'everything' install and how to get the chroot flavor installed.
Also, there is a general-interest part. How many people install private networks in such a way that essentially every local connection will send a reverse-DNS query for their private space to the internet root nameservers? The people who do that aren't likely to be reading the bind mailing list...
} } There was a distribution-packaging part of this discussion that wasn't } answered. If someone wants to bring it back on track they could explain } what you get with an 'everything' install and how to get the chroot } flavor installed. } } Also, there is a general-interest part. How many people install } private networks in such a way that essentially every local connection } will send a reverse-DNS query for their private space to the } internet root nameservers? The people who do that aren't likely } to be reading the bind mailing list... } } -- } Les Mikesell
Les,
i started to answer it and then thought better. i know what you asked and mean yet i just trial and error till i found what works for me re:
rpm -qa | grep whatever yum whatprovides yum install yum remove
etc and my servers among the "other" things you are asking or implying.
yet, i have setup a lot of linux boxes over the years from slackware 0.9x thru various Redhat and others and now CentOS etc.
maybe you should ask again to the CentOS devel people???
regards,
- rh
-- Robert Hanson Abba Communications http://www.abbacomm.net
On Thu, 2005-08-11 at 10:40, Robert Hanson wrote:
} } There was a distribution-packaging part of this discussion that wasn't } answered. If someone wants to bring it back on track they could explain } what you get with an 'everything' install and how to get the chroot } flavor installed.
i started to answer it and then thought better. i know what you asked and mean yet i just trial and error till i found what works for me re:
rpm -qa | grep whatever yum whatprovides yum install yum remove
etc and my servers among the "other" things you are asking or implying.
Same here, and unfortunately I take the same approach to reading the documentation, just hunting for details when I need them. I was hoping that the package groupings and a way to browse them was documented somewhere and I had just missed it. You can use synaptic to wade through the individual packages, but where do you find the overview/detail of what you get in the groupings?
I am intently trying to avoid sending anything to the root servers w/o checking with my Astaro firewall first. Hence why i started with a caching name server. Now i am looking to setup a full internal DNS server for all of my network since i intend to implement AD in the future(again for learning purposes)
Les Mikesell wrote:
On Thu, 2005-08-11 at 08:55, Lamar Owen wrote:
On Thursday 11 August 2005 01:52, Jesse wrote:
Folks. I thought this was a CentOS list not a Bind list... This list is getting out of hand lately.
You know, in a way I agree with this sentiment.
However, at the same time I also see that discussing the CentOS distribution of bind is not a big deal for this list.
There was a distribution-packaging part of this discussion that wasn't answered. If someone wants to bring it back on track they could explain what you get with an 'everything' install and how to get the chroot flavor installed.
Also, there is a general-interest part. How many people install private networks in such a way that essentially every local connection will send a reverse-DNS query for their private space to the internet root nameservers? The people who do that aren't likely to be reading the bind mailing list...
On Thu, 2005-08-11 at 10:41, William Warren wrote:
I am intently trying to avoid sending anything to the root servers w/o checking with my Astaro firewall first.
It didn't seem to be answering for your 192.168.0.200 request. What makes you think it isn't sending those revere lookups out?
<chuckle> it can't right now since it's not running..:) I intend to configure it that way.
I got it reinstalled chrooted thanks David's earlier post. Now i have a fresh and clean bind server with no zones defined at all. Time to figure out how to set it up(using webmin due to another suggestion on the list) and ensure it's only place to go for information is the Astaro machine. I am going to define all my internal machines since they are DHCP static by MAC..:)
Les Mikesell wrote:
On Thu, 2005-08-11 at 10:41, William Warren wrote:
I am intently trying to avoid sending anything to the root servers w/o checking with my Astaro firewall first.
It didn't seem to be answering for your 192.168.0.200 request. What makes you think it isn't sending those revere lookups out?
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
On 8/11/05, David Evennou de@data-masters.com wrote:
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
You can try installing GRUB to the MBR.
Moreover have you partitioned your system as /dev/hda2 as the /boot partition? if yes, then there is a different method for dual booting.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
I installed GRUB to the MBR on a previous installation, and that wiped out the XP boot capability. Yes, /dev/hda2 is the CentOS boot partition.
Thanks, David
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:05 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/11/05, David Evennou de@data-masters.com wrote:
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS
Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
You can try installing GRUB to the MBR.
Moreover have you partitioned your system as /dev/hda2 as the /boot
partition?
if yes, then there is a different method for dual booting.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 8/12/05, David Evennou de@data-masters.com wrote:
I installed GRUB to the MBR on a previous installation, and that wiped out the XP boot capability. Yes, /dev/hda2 is the CentOS boot partition.
Thanks, David
If your Windows XP Partions are formatted as NTFS, then Cent OS can't detect it. (The Cent OS kernel lacks NTFS modules.) So, the final result is that you can't boot to windows XP, but cent OS works fine. (I can provide you some more information on this, if you are willing to do it.)
Moreover, you can still get that GRUB installed to the MBR, then get the Cent OS working, and then patch the Kernel to include NTFS Modules. Then add the required lines to the /boot/grub/menu.lst file and get windows working. (I can provide you some more information on this, if you are willing to do it.)
I am sorry, cuz I don't know how to use the /boot partition properly. I mean, I am not aware how to make the /boot partition to work, by installing GRUB on it.
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:05 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/11/05, David Evennou de@data-masters.com wrote:
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS
Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
You can try installing GRUB to the MBR.
Moreover have you partitioned your system as /dev/hda2 as the /boot
partition?
if yes, then there is a different method for dual booting.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Side note: I have a Ubuntu Linux installation the was installed on a XP box. I followed the following instructions, both boot options, Linux and XP were available choices.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Install GRUB on the first sector of the /boot partition. DO NOT INSTALL IT ON THE MBR!. If you are performing the Red Hat installation, for the "Boot Loader Installation" screen: a.. Select "Use GRUB as the boot loader" b.. Select Install Boot Loader record on "...First sector of boot partition". c.. After finishing the Red Hat installation, reboot into Linux. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I thought the same procedure would work, because the CentOS Installation gave me the choice of where to install GRUB using the "Advanced" setting.
Note: the information above is from this link:
http://www.geocities.com/epark/linux/grub-w2k-HOWTO.html
Thanks, David
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:16 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/12/05, David Evennou de@data-masters.com wrote:
I installed GRUB to the MBR on a previous installation, and that wiped
out
the XP boot capability. Yes, /dev/hda2 is the CentOS boot partition.
Thanks, David
If your Windows XP Partions are formatted as NTFS, then Cent OS can't detect it. (The Cent OS kernel lacks NTFS modules.) So, the final result is that you can't boot to windows XP, but cent OS works fine. (I can provide you some more information on this, if you are willing to do
it.)
Moreover, you can still get that GRUB installed to the MBR, then get the Cent OS working, and then patch the Kernel to include NTFS Modules. Then add the required lines to the /boot/grub/menu.lst file and get windows working. (I can provide you some more information on this, if you are willing to do
it.)
I am sorry, cuz I don't know how to use the /boot partition properly. I mean, I am not aware how to make the /boot partition to work, by installing GRUB on it.
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:05 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/11/05, David Evennou de@data-masters.com wrote:
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS
Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
You can try installing GRUB to the MBR.
Moreover have you partitioned your system as /dev/hda2 as the /boot
partition?
if yes, then there is a different method for dual booting.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Aug 12, 2005, at 9:44 AM, David Evennou wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Add these lines to your /etc/grub.conf
title Windows XP rootnoverify (hd0,1) chainloader +1
making sure to substitute the right partition name for (hd0,1)
On 8/12/05, Tony Schreiner schreian@bc.edu wrote:
On Aug 12, 2005, at 9:44 AM, David Evennou wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Add these lines to your /etc/grub.conf
title Windows XP rootnoverify (hd0,1) chainloader +1
making sure to substitute the right partition name for (hd0,1)
Doing this alone won't help, as the Cent OS kernel doesn't have support for NTFS File System. (Which Windows XP Uses by default)
I think following the instructions in my previous post should help.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Fri, Aug 12, 2005 at 02:09:25PM +0000, duffmckagan enlightened us:
On Aug 12, 2005, at 9:44 AM, David Evennou wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Add these lines to your /etc/grub.conf
title Windows XP rootnoverify (hd0,1) chainloader +1
making sure to substitute the right partition name for (hd0,1)
Doing this alone won't help, as the Cent OS kernel doesn't have support for NTFS File System. (Which Windows XP Uses by default)
I think following the instructions in my previous post should help.
This is misinformation. The kernel only needs to know about filesystems it uses. At boot time, the kernel isn't even loaded, so in order to *boot* an NTFS, all you need to do is add the above to your grub.conf. If you want to access your NTFS partition from Linux, then yes, your kernel will need NTFS support.
Matt
On 8/12/05, Matt Hyclak hyclak@math.ohiou.edu wrote:
On Fri, Aug 12, 2005 at 02:09:25PM +0000, duffmckagan enlightened us:
On Aug 12, 2005, at 9:44 AM, David Evennou wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Add these lines to your /etc/grub.conf
title Windows XP rootnoverify (hd0,1) chainloader +1
making sure to substitute the right partition name for (hd0,1)
Doing this alone won't help, as the Cent OS kernel doesn't have support for NTFS File System. (Which Windows XP Uses by default)
I think following the instructions in my previous post should help.
This is misinformation. The kernel only needs to know about filesystems it uses. At boot time, the kernel isn't even loaded, so in order to *boot* an NTFS, all you need to do is add the above to your grub.conf. If you want to access your NTFS partition from Linux, then yes, your kernel will need NTFS support.
Matt
LOL....okay. That is true. I regret my mistake.
Well, the reason for Cent OS not detecting the NTFS partition, and not including Windows in GRUB Menu, is that the Kernel doesnt' support NTFS.
Sorry again. But, I guess that the NTFS Modules would be needed if David wants to access the NTFS partitions from Linux.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 12/08/2005 15:19, duffmckagan wrote:
Well, the reason for Cent OS not detecting the NTFS partition, and not including Windows in GRUB Menu, is that the Kernel doesnt' support NTFS.
That's not true, either. I have installed CentOS (both 3 and 4) on a number of machines with NTFS partitions and it has detected and added a GRUB menu entry just fine.
Mike.
This message has been scanned for viruses by MailController - www.MailController.altohiway.com
Mike,
I am a relatively new Linux convert, about 8 months.
Can you please detail your procedure, step by step?
Thanks, David
----- Original Message ----- From: "Mike Zanker" mike@zanker.org To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 10:33 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 12/08/2005 15:19, duffmckagan wrote:
Well, the reason for Cent OS not detecting the NTFS partition, and not including Windows in GRUB Menu, is that the Kernel doesnt' support NTFS.
That's not true, either. I have installed CentOS (both 3 and 4) on a number of machines with NTFS partitions and it has detected and added a GRUB menu entry just fine.
Mike.
This message has been scanned for viruses by MailController -
www.MailController.altohiway.com
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 12/8/05 17:52, David Evennou (Data Masters) wrote:
I am a relatively new Linux convert, about 8 months.
Can you please detail your procedure, step by step?
There is no procedure - I meant that CentOS detects and adds a menu entry during installation.
Others have stated the procedure for an installed system.
Mike.
This message has been scanned for viruses by MailController - www.MailController.altohiway.com
Interesting!
I did that procedure earlier and did not have a choice for booting XP.
Did you pick the defaults for installing GRUB?
Thanks, David
----- Original Message ----- From: "Mike Zanker" mike@zanker.org To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 12:58 PM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 12/8/05 17:52, David Evennou (Data Masters) wrote:
I am a relatively new Linux convert, about 8 months.
Can you please detail your procedure, step by step?
There is no procedure - I meant that CentOS detects and adds a menu entry during installation.
Others have stated the procedure for an installed system.
Mike.
This message has been scanned for viruses by MailController -
www.MailController.altohiway.com
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Fri, 2005-08-12 at 15:33 +0100, Mike Zanker wrote:
On 12/08/2005 15:19, duffmckagan wrote:
Well, the reason for Cent OS not detecting the NTFS partition, and not including Windows in GRUB Menu, is that the Kernel doesnt' support NTFS.
The centosplus "unsupported" kernels do support NTFS.
That's not true, either. I have installed CentOS (both 3 and 4) on a number of machines with NTFS partitions and it has detected and added a GRUB menu entry just fine.
Same here - windows partitions have, in my experience, always (well as least since solid GRUB support was added - but we're not talking LILO here) been detected and added by grub on RH-derived systems. [Not so for detecting/adding other Linux installations, unfortunately, but that's not too hard to do manually if you're like me and have 4 or 5 different OSs installed on "play" machines - disk space is cheap these days.]
Phil
On Fri, 2005-08-12 at 14:09 +0000, duffmckagan wrote:
On 8/12/05, Tony Schreiner schreian@bc.edu wrote:
On Aug 12, 2005, at 9:44 AM, David Evennou wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Add these lines to your /etc/grub.conf
title Windows XP rootnoverify (hd0,1) chainloader +1
making sure to substitute the right partition name for (hd0,1)
The above should work ... you may need to play with the (hd0,1) line ... I had to after creating a FAT32 partition to share data between Windows 2000 and Centos 4.1.
Doing this alone won't help, as the Cent OS kernel doesn't have support for NTFS File System. (Which Windows XP Uses by default)
I think following the instructions in my previous post should help.
The kernel does not need NTFS support ... booting to Win2K/XP only requires that the boot loader be able to find the partition.
Regards, Paul
On 8/12/05, David Evennou de@data-masters.com wrote:
OK, thanks!
Yes, I would like the information to get Windows XP running after installing GRUB on the MBR.
Side note: I have a Ubuntu Linux installation the was installed on a XP box. I followed the following instructions, both boot options, Linux and XP were available choices.
Install GRUB on the first sector of the /boot partition. DO NOT INSTALL IT ON THE MBR!. If you are performing the Red Hat installation, for the "Boot Loader Installation" screen: a.. Select "Use GRUB as the boot loader" b.. Select Install Boot Loader record on "...First sector of boot partition". c.. After finishing the Red Hat installation, reboot into Linux.I thought the same procedure would work, because the CentOS Installation gave me the choice of where to install GRUB using the "Advanced" setting.
Note: the information above is from this link:
Yeah. I know the above article. and I have read that thing already, but never bothered to try that out.
What you can do is...simply install Cent OS, and Windows XP in dual boot configuration. Install the GRUB to the MBR.
Sorry, but i can't really help you with the /boot partition. (I would prefer to remove it, and make two partitions / and /home.) Now, when Cent OS is completely installed, take a look at the following link, download the Correct RPM for your Kernel, and run it as root
http://linux-ntfs.sourceforge.net/rpm/redhat9.html
root@computer#rpm -ivh <kernel module you have downloaded>
After that, edit your /boot/grub/menu.lst file to include the following. (you can simply append this text )
title Windows XP rootnoverify (hd0,0) makeactive chainloader +1
Hope all this helps.
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:16 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/12/05, David Evennou de@data-masters.com wrote:
I installed GRUB to the MBR on a previous installation, and that wiped
out
the XP boot capability. Yes, /dev/hda2 is the CentOS boot partition.
Thanks, David
If your Windows XP Partions are formatted as NTFS, then Cent OS can't detect it. (The Cent OS kernel lacks NTFS modules.) So, the final result is that you can't boot to windows XP, but cent OS works fine. (I can provide you some more information on this, if you are willing to do
it.)
Moreover, you can still get that GRUB installed to the MBR, then get the Cent OS working, and then patch the Kernel to include NTFS Modules. Then add the required lines to the /boot/grub/menu.lst file and get windows working. (I can provide you some more information on this, if you are willing to do
it.)
I am sorry, cuz I don't know how to use the /boot partition properly. I mean, I am not aware how to make the /boot partition to work, by installing GRUB on it.
----- Original Message ----- From: "duffmckagan" mckagan@gmail.com To: "CentOS mailing list" centos@centos.org Sent: Friday, August 12, 2005 9:05 AM Subject: Re: [CentOS] Need Dual Boot Installation Help
On 8/11/05, David Evennou de@data-masters.com wrote:
During CentOS 4.1 Installation, I selected Advanced->GRUB Boot Loader->Install to /dev/hda2. (the /dev/hda1 partition as XP on it)
After reboot, I still only get XP booted and no choices for CentOS
Linux.
I tried GRUB-INSTALL /dev/hda2 at the boot prompt, but that does not execute.
What am I missing?
TIA, David
You can try installing GRUB to the MBR.
Moreover have you partitioned your system as /dev/hda2 as the /boot
partition?
if yes, then there is a different method for dual booting.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-- "No-one dies a virgin. Life screws everyone." _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Friday 12 August 2005 09:16, duffmckagan wrote:
On 8/12/05, David Evennou de@data-masters.com wrote:
I installed GRUB to the MBR on a previous installation, and that wiped out the XP boot capability. Yes, /dev/hda2 is the CentOS boot partition.
If your Windows XP Partions are formatted as NTFS, then Cent OS can't detect it. (The Cent OS kernel lacks NTFS modules.) So, the final result is that you can't boot to windows XP, but cent OS works fine. (I can provide you some more information on this, if you are willing to do it.)
This is incorrect. I am dual-booting my Dell Laptop right now; see: [root@localhost lowen]# /sbin/fdisk -l
Disk /dev/hda: 60.0 GB, 60011642880 bytes 255 heads, 63 sectors/track, 7296 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System /dev/hda1 1 5 40131 de Dell Utility /dev/hda2 * 6 1049 8385930 7 HPFS/NTFS /dev/hda3 1050 1065 128520 83 Linux /dev/hda4 1066 7296 50050507+ f W95 Ext'd (LBA) /dev/hda5 1066 2370 10482381 83 Linux /dev/hda6 2371 3135 6144831 83 Linux /dev/hda7 3136 3266 1052226 82 Linux swap /dev/hda8 3267 5877 20972826 b W95 FAT32 /dev/hda9 5878 7296 11398086 83 Linux [root@localhost lowen]#
As you can see, WinXP on NTFS. Now: [root@localhost lowen]# cat /boot/grub/grub.conf # grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You have a /boot partition. This means that # all kernel and initrd paths are relative to /boot/, eg. # root (hd0,2) # kernel /vmlinuz-version ro root=/dev/hda6 # initrd /initrd-version.img #boot=/dev/hda default=0 timeout=5 splashimage=(hd0,2)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.9-11.106.unsupported) root (hd0,2) kernel /vmlinuz-2.6.9-11.106.unsupported ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.9-11.106.unsupported.img title CentOS (2.6.9-5.0.5.106.unsupported) root (hd0,2) kernel /vmlinuz-2.6.9-5.0.5.106.unsupported ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.9-5.0.5.106.unsupported.img title CentOS (2.6.9-5.0.5.EL) root (hd0,2) kernel /vmlinuz-2.6.9-5.0.5.EL ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.9-5.0.5.EL.img title CentOS 4.0 (2.6.9-5.0.3.EL) root (hd0,2) kernel /vmlinuz-2.6.9-5.0.3.EL ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.9-5.0.3.EL.img title Other rootnoverify (hd0,1) chainloader +1 [root@localhost lowen]#
Boot target 'Other' is WinXP, and it Just Works and Just Worked after installation.
What one has to be careful about is incorrect BIOS disk geometry reporting and geometry differences between what BIOS, GRUB, and the Linux kernel believe. My Dell Inspiron 600m Just Works because it has a decent BIOS and I knew how to properly set up my partitions. But other than that no NTFS modules are required.
I set up GRUB to be installed in the MBR, and it Just Worked.
duffmckagan mckagan@gmail.com wrote:
Moreover, you can still get that GRUB installed to the MBR, then get the Cent OS working, and then patch the Kernel to include NTFS Modules. Then add the required lines to the /boot/grub/menu.lst file and get windows working.
You do _not_ need NTFS support in GRUB to boot Windows. GRUB will merely use the chainloader to pass off boot to the NTFS partition, even though it can't read it.
BTW, note there are _issues_ with GRUB and Windows NT5.1 (XP) Service Packs / Hotfixes. It seems that on its bootable drive (what MS calls the "System Drive") on a "Basic Disc" (legacy BIOS/DOS Primary/Extended/Logical disk slicing), it uses portions of the MBR to store "hidden info." It will overwrite portions of GRUB. If you re-install GRUB, it will overwrite portions of these hidden areas, and XP SP1+hostfixes or SP2[+hotfixes] will not boot.
The conditions seem to vary from system to system, but I have run into it regularly now -- especially with many SP1 hotfixes and definitely with SP2. There also seem to be some conflicting hotfixes, especially SP2+, that constantly change things (even on a 100% Windows XP system). If I had to guess, and based on my views of the MBR with a hex editor, I think Microsoft is still trying to be compliant with the new ATA48 spec (lont story).
But _regardless_, I find that by keeping my "system drive" (C: filesystem) with the NTLDR, BOOT.INI, etc... XP partition within the first 32GiB (at least when geometry is <4096 cylinders, =255 heads and =63 sectors) of the hard drive will typically work.
I literally had a situation where XP SP1 was booting fine in the first 64GiB, then I loaded some hotfixes, and it refused to boot. Sure enough, upon inspection of the "damaged" (non-GRUB booting) MBR (which I made a copy of before re-installing, and then losing the ability of chainloader to work), I noticed the same bytes as SP2.
On 8/12/05, Bryan J. Smith b.j.smith@ieee.org wrote:
duffmckagan mckagan@gmail.com wrote:
Moreover, you can still get that GRUB installed to the MBR, then get the Cent OS working, and then patch the Kernel to include NTFS Modules. Then add the required lines to the /boot/grub/menu.lst file and get windows working.
You do _not_ need NTFS support in GRUB to boot Windows. GRUB will merely use the chainloader to pass off boot to the NTFS partition, even though it can't read it.
BTW, note there are _issues_ with GRUB and Windows NT5.1 (XP) Service Packs / Hotfixes. It seems that on its bootable drive (what MS calls the "System Drive") on a "Basic Disc" (legacy BIOS/DOS Primary/Extended/Logical disk slicing), it uses portions of the MBR to store "hidden info." It will overwrite portions of GRUB. If you re-install GRUB, it will overwrite portions of these hidden areas, and XP SP1+hostfixes or SP2[+hotfixes] will not boot.
The conditions seem to vary from system to system, but I have run into it regularly now -- especially with many SP1 hotfixes and definitely with SP2. There also seem to be some conflicting hotfixes, especially SP2+, that constantly change things (even on a 100% Windows XP system). If I had to guess, and based on my views of the MBR with a hex editor, I think Microsoft is still trying to be compliant with the new ATA48 spec (lont story).
But _regardless_, I find that by keeping my "system drive" (C: filesystem) with the NTLDR, BOOT.INI, etc... XP partition within the first 32GiB (at least when geometry is <4096 cylinders, =255 heads and =63 sectors) of the hard drive will typically work.
I literally had a situation where XP SP1 was booting fine in the first 64GiB, then I loaded some hotfixes, and it refused to boot. Sure enough, upon inspection of the "damaged" (non-GRUB booting) MBR (which I made a copy of before re-installing, and then losing the ability of chainloader to work), I noticed the same bytes as SP2.
-- Bryan J. Smith | Sent from Yahoo Mail mailto:b.j.smith@ieee.org | (please excuse any http://thebs413.blogspot.com/ | missing headers)
Thanks for the information, and removing away my misconception. This mailing list rocks. :)
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
duffmckagan mckagan@gmail.com wrote:
Thanks for the information, and removing away my misconception. This mailing list rocks. :)
I recently did a (quickly hacked together) set of presentations for the St. Louis UNIX User's Group which is local to my current client. They cover a _lot_ of the "low-level" interoperability issues between Linux and Windows, almost all of which are inherent Windows issues (e.g., issues with NTFS' design).
Probably of greatest interests are Part I which gives some PC basic disk geometry/slicing intro/coverage in the mid-to-latter slides:
http://www.geocities.com/thebs413/SLUUG_LowLevelInterop_Part1.pdf
As well as Part II that talks about Linux's 100% standards compliance to ATA and BIOS Extended Interrupt 13h Disk Services and various issues "buggy BIOSes" (which are still commonplace *COUGH*IBM*COUGH*) plus Microsoft's _lack_ of standards in the early slides:
http://www.geocities.com/thebs413/SLUUG_LowLevelInterop_Part2.pdf
Although you might want to read Part III which also covers more about why NTFS is such an interoperability pain for Windows NT/2000/XP systems themselves here:
http://www.geocities.com/thebs413/SLUUG_LowLevelInterop_Part3.pdf
Top-part: I've been purposely holding off commenting to see what others would say. With that said, I'm going to make a "summary" of "things to know" for handling both private and public DNS for an organization -- SOHO or bigger.
William Warren hescominsoon@emmanuelcomputerconsulting.com wrote:
I am intently trying to avoid sending anything to the root servers w/o checking with my Astaro firewall first. Hence why i started with a caching name server. Now i am looking to setup a full internal DNS server for
all
of my network
Correct. You should do this because your internal systems might be trying to query public servers for internal-only names/addresses.
To summarize, bullet-by-bullet, of what most people have already said and most Enterprises do:
1. Use a DNS domain that is _unique_ for the private network 1.a. If you already have a domain name for your company/organization, create the private network as a _subdomain_ -- _never_ call it the same. 1.b. If you are using ADS-integrated DNS, consider making ADS domains their own subdomains -- or even a different domain with an _invalid_ TLD (e.g., ads.mycompany.prvtld) so it is _never_ resolvable on the internet.
2. Setup at least 2 DNS servers, one primary, one secondary 2.a. The Primary DNS has both a forward and reverse zone for the private network 2.b. Both Primary and Secondary DNS have "forwarder" entries to fetch public names 2.c. NOTE: While you can have multiple DNS domains to a subnet, but only 1 reverse zone -- so this might be an issue with separate ADS-integrated and BIND DNS servers, just FYI.
3. Set the DNS servers of all Windows and UNIX clients to _only_ the _internal_ DNS servers 3.a. _Never_ allow private LAN nodes to query public DNS servers, only the internal DNS servers 3.b. The "forwarder" entries on the internal DNS servers handle fetching public DNS names/IPs
4. Block any outgoing UDP/TCP traffic to destination port 53 (if you're not blocking all outgoing by default already), except for the internal DNS servers
5. Hack the registry on Windows NT5+ (2000+) so they do not cache DNS resolution failures (as we previously discussed)
Now if you are going to want your public IP addresses to be reachable, it depends on if you have dynamic or static addresses. For dynamic, use a service that will map your name/IP upon update. If you have a block of static IPs, then do the following ...
6. Again, use _separate_ domains for private and public 6.a. I typically _still_ use a subdomain for the public (e.g., public.company.com) 6.b. I then use CNAMEs from the "root" domain (e.g., mycompany.com) so I _never_ have to change the "root" domain DNS (and can leave it "read only").
7. If your public IP block is not a perfect class A, B or C -- you can_not_ maintain your reverse zones directly. 7.a. It's very typical to have a "classless" IP block of only 4 (/30), 8 (/29), 16 (/28) or 32 (/27) addresses 7.b. This could be that your ISP with the "full" class C handle it 7.c. Alternatively, per RFC2317 "Classless IN-ADDR.ARPA delegation" ( http://www.faqs.org/rfcs/rfc2317.html ), you setup a reverse zone per your ISP's format. E.g., if your ISp gives you 100.150.200.192-207, 16 addresses, 255.255.255.240 (/28 CIDR format) subnet mask, your $ORIGIN for your reverse zone would typically be (there are alternatives to this, depending on your ISP's entry): $ORIGIN 192/28.200.150.100.in-addr.arpa. ^^^ ^^ ^^^ ^^^ ^^^ D E C B A A, B and C = 1st, 2nd and 3rd octects, respectively D = start of 4th octet block E = CIDR format subnet mask (28 = 255.255.255.240)
since i intend to implement AD in the future(again for learning purposes)
ADS can use DNS in various modes.
One proprietary one is "ADS-integrated" whereby ADS provides peer-replicated DNS (and BIND can _only_ be secondary to it).
There are a few 3rd party solutions which also offer a unified layer-2/3 DHCP/DDNS/DNS/etc... name resolution framework with peer replication, but they are not free (let alone _none_ are open source).
Frankly, although there are a few "wrapper management tools" for enterprises around ISC DHCP and BIND, I'd really like to see an "unified" layer-2/3 name service solution come from the open source world. God knows with IPv6, it is also redundant to keep DHCP and DNS separate, and not under the control of a single daemon.
"Bryan J. Smith" b.j.smith@ieee.org wrote:
Frankly, although there are a few "wrapper management tools" for enterprises around ISC DHCP and BIND, I'd really like to see an "unified" layer-2/3 name service solution come from the open source world. God knows with IPv6, it is also redundant to keep DHCP and DNS separate, and not under the control of a single daemon.
Er, let me rephrase that, a single _set_ of daemons that are drive by the same framework/config files. In other words, daemons that worked directly with each other.
E.g., a good set of daemons would include: - RARP protocol (BOOTP/DHCP) - DDNS updates (from clients, including "override" logic) - DNS queries - DNS zone transfers (legacy BIND/DNS secondary servers) - Legacy ident/proxy/service integration (NetBIOS/WINS, SAP/NLSP) - Peer replication (to the same daemons on other systems)
Especially since IPv6 uses MAC addresses, and any standard node can have up to three (3) IPv6 addresses.
William Warren wrote:
I highly appreciate everyone's help in the caching name server bit. Building on that one:
Everyone knows my issues that i am having in terms of reverse lookup so the windows clients won't freak. I have decided to switch focus..as it seemed to make sense. I am now looking to setup a full DNS server for my internal network. Everything is actually set statically via DHCP using MAC addresses.(see attached jpeg).
Do i have to manually enter every machine or can the dns populate the names and ips via dhcp?
you will have to setup keys between dhcpd and named for the zone(s) that will allow dynamic updates.
-
Bryan J. Smith -
David Evennou -
David Evennou (Data Masters)
-
duffmckagan -
Feizhou -
Jesse -
Lamar Owen -
Les Mikesell -
Matt Hyclak -
Mike Zanker -
Paul -
Phil Schaffner -
Robert Hanson -
Tony Schreiner -
William Warren