I subscribed, and am getting emails hand over fist. They'll bust my inbox quota if I miss a day.
I went to the web site for my subscription and tried to disable delivery; I'll follow this list on gmane or not at all -- and that works with every other mailman list I've tried.
It wouldn't let me -- kept reverting to the password screen; I kept giving it the password, and making the change; and next time, the change was not there -- even though my password was letting me in.
I sent an email to the address specified on the subscription website, asking for help. It sent me a piece of boilerplate in reply telling me I can't post (wanna bet?) -- because I'm not subscribed!
This is a high traffic mailing list, thus it's very normal to get many emails in a days.
It's easy to unsubscribe.
1. Go to http://lists.centos.org/mailman/listinfo/centos 2. Enter you e-mail in the box at the bottom and press "unsubscribe or edit" 3. Enter your password to continue, or press "remind" to have your password mailed to you. 4. In the large screen select "/Yes, I really want to unsubscribe" and press Unsubscribe
-- sukru
/beartooth wrote:
I subscribed, and am getting emails hand over fist. They'll bust my inbox quota if I miss a day.
I went to the web site for my subscription and tried to disable delivery; I'll follow this list on gmane or not at all -- and that works with every other mailman list I've tried.
It wouldn't let me -- kept reverting to the password screen; I kept giving it the password, and making the change; and next time, the change was not there -- even though my password was letting me in.
I sent an email to the address specified on the subscription website, asking for help. It sent me a piece of boilerplate in reply telling me I can't post (wanna bet?) -- because I'm not subscribed!
On Thursday 23 June 2005 08:58, Sukru TIKVES wrote:
This is a high traffic mailing list, thus it's very normal to get many emails in a days.
Slightly off topic, but this is not a high traffic mailing list. Try fedora-list if you want something with a little more traffic.
I get over 1500 e-mails per day. This list has a low volume. :-)
Lamar Owen wrote:
On Thursday 23 June 2005 08:58, Sukru TIKVES wrote:
This is a high traffic mailing list, thus it's very normal to get many emails in a days.
Slightly off topic, but this is not a high traffic mailing list. Try fedora-list if you want something with a little more traffic.
I get over 1500 e-mails per day. This list has a low volume. :-)
I found out that using gmane/nntp/thunderbird is the best way, when practical, to access mailing list. Instant acces to the archives, easy threaded view, only downloading headers, downloads messages only if read.
On Thu, 23 Jun 2005, Ugo Bellavance wrote:
I found out that using gmane/nntp/thunderbird is the best way, when practical, to access mailing list. Instant acces to the archives, easy threaded view, only downloading headers, downloads messages only if read.
Gmane is what I'm trying to do, as I thought I had said -- though I stick to Pan.
On Thu, 23 Jun 2005, Lamar Owen wrote:
Slightly off topic, but this is not a high traffic mailing list. Try fedora-list if you want something with a little more traffic.
I get over 1500 e-mails per day. This list has a low volume. :-)
Slightly more off topic (because that's not the issue) : yaaa, shuuure -- if you have a .edu address. Been there done that. Try it from adelphia the utterly accursed.
On Jun 23, 2005, at 12:25 PM, Bear Tooth wrote:
Slightly more off topic (because that's not the issue) : yaaa, shuuure -- if you have a .edu address. Been there done that. Try it from adelphia the utterly accursed.
sounds like someone needs to set up his own mail server :) weren't we talking about this just a few days ago?
-steve (hoping to avoid provoking more qmail/postfix sniping)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Thu, Jun 23, 2005 at 12:38:08PM -0400, Steve Huff wrote:
On Jun 23, 2005, at 12:25 PM, Bear Tooth wrote:
Slightly more off topic (because that's not the issue) : yaaa, shuuure -- if you have a .edu address. Been there done that. Try it from adelphia the utterly accursed.
sounds like someone needs to set up his own mail server :) weren't we talking about this just a few days ago?
-steve (hoping to avoid provoking more qmail/postfix sniping)
Does that mean I'm free to advocate for Exim ? :)
[]s
- -- Rodrigo Barbosa rodrigob@suespammers.org "Quid quid Latine dictum sit, altum viditur" "Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)
On Thursday 23 June 2005 12:25, Bear Tooth wrote:
On Thu, 23 Jun 2005, Lamar Owen wrote:
Slightly off topic, but this is not a high traffic mailing list. Try fedora-list if you want something with a little more traffic.
I get over 1500 e-mails per day. This list has a low volume. :-)
Slightly more off topic (because that's not the issue) : yaaa, shuuure -- if you have a .edu address. Been there done that. Try it from adelphia the utterly accursed.
Steering this towards something that is at least slightly on topic, I run my own e-mail servers on CentOS4 (using the easily understood sendmail; easily understood compared to what I used 14 years ago when I first 'registered' lorc.uucp in the Bad Old Days of dialup uucp).
I have, let's see, 24 e-mail addresses, which includes postmaster@pari.edu as well as postmaster at a couple of .com's, three .org's, and a handful of .net's. The pari.edu mailserver is not currently on CentOS, though; it's on Aurora 1.92+ on an UltraSPARC. Thought about putting it on AlphaCore on the big quad AlphaServer downstairs; but the 250MHz USPARCII runs circles around the quad 275MHz EV45 in the AS2100, at least when the big CPU hog is MailScanner and cronies. For lamarowen.net, which IS running CentOS 4, I use a Soyo DRAGON and a cheap Duron with 768MB of RAM and 800GB of drives. Small box, under my desk. Makes a nice footwarmer. For another site, a Dell PowerEdge 750 (which has some issues with the release CentOS 4 kernel that the U1 kernel seems to have fixed) with sendmail, amavisd-new, clamav, and spamassassin (milter mode) works like a charm. For yet another site the same soup on a PE1750 does the trick; CentOS 4 again. Good stuff.
My ISP e-mail account gets a mere couple dozen e-mails per day, and has a very large inbox. Back in 1996 when that ISP first went into business, all my list e-mail went there, and I got up to over 2,000 e-mails per day. Their OpenVMS servers had large inboxes, and things worked fine.
On Thu, 23 Jun 2005 13:31:35 -0400, Lamar Owen wrote:
Steering this towards something that is at least slightly on topic, I run my own e-mail servers on CentOS4 (using the easily understood sendmail; easily understood compared to what I used 14 years ago when I first 'registered' lorc.uucp in the Bad Old Days of dialup uucp).
I have, let's see, 24 e-mail addresses, [....]
You, Sir!, are obviously a real dyed in the wool, simon pure quill, hundred percent genuine Alpha Double Plus Technoid -- and a connoisseur of hardware to boot. I salute you!
I will watch for advice from you in particular if I do decide to switch to CentOS, most especially about a mailserver; once I manage to network my machines at all, I'd like to set one up.
As for me, I don't even speak hardware. I ran RedHat 6.0 through 9 before Fedora, by the seat of my pants and the help of the Net.
I'm on this list because I'm getting near the end of my rope after a good dozen installs, upgrades, and re-installs just since Fedora first came out. It changes too fast for me -- in fine and admirable ways, but I can't keep up. I'll install FC4 when the installation media get here, and that may be it.
So, getting back to the real topic, I'm especially interested in matters of stability, if that's the term of art: CentOS is basically a clone of RHEL without the price nor the phone support -- right? I ran boughten RH, and had RHN all those years, and never picked up a phone. RedHat's updates, and help from online, kept it on an even keel.
Gretchenfrage : is it looking like an electronically ignorant old retired fart will be able to install CentOS, set each machine to do nightly yum update, and go back to doing what he does, maybe for a year or two at a time? The learning curve at least ought to be minimal, for a RedHat refugee ...
Am Fr, den 24.06.2005 schrieb beartooth um 19:02:
I'm on this list because I'm getting near the end of my rope after a good dozen installs, upgrades, and re-installs just since Fedora first came out. It changes too fast for me -- in fine and admirable ways, but I can't keep up. I'll install FC4 when the installation media get here, and that may be it.
Yes, if you are not the enthusiastic Linux user, requesting to have always the latest, biggest, most promising software version - [b]le[a,e]ding edge - then CentOS is certainly the better OS for you.
So, getting back to the real topic, I'm especially interested in matters of stability, if that's the term of art: CentOS is basically a clone of RHEL without the price nor the phone support -- right? I ran boughten RH, and had RHN all those years, and never picked up a phone. RedHat's updates, and help from online, kept it on an even keel.
From my experience CentOS4 is very solid and a good choice for server systems. Updates come in time and for free.
Gretchenfrage : is it looking like an electronically ignorant old retired fart will be able to install CentOS, set each machine to do nightly yum update, and go back to doing what he does, maybe for a year or two at a time? The learning curve at least ought to be minimal, for a RedHat refugee ...
Though I am neither Doc Faust, nor have a relation with Mephistopheles, I am sure you have more pleasure running CentOS in the way you intend than you had from your Fedora experiences so far (knowing some of your postings on the Fedora and the Fedora Legacy Project user lists). The only thing which I generally do not recommend is running automatic nightly yum updates. It is little administration to do updates manually and you always know what happens by watching at yum's activity output.
Alexander
beartooth wrote:
On Thu, 23 Jun 2005 13:31:35 -0400, Lamar Owen wrote:
So, getting back to the real topic, I'm especially interested in matters of stability, if that's the term of art: CentOS is basically a clone of RHEL without the price nor the phone support -- right?
To me, yes. I run mail filtering servers (MailScanner) based on CentOS, a web development/backup/multi-purpose server, and a production php/mysql web app server with CentOS and they're doing great. I had same stability with Tao Linux before, but I switched to CentOS a while ago.
I ran boughten RH, and had RHN all those years, and never picked up a phone. RedHat's updates, and help from online, kept it on an even keel.
Gretchenfrage : is it looking like an electronically ignorant old retired fart will be able to install CentOS, set each machine to do nightly yum update, and go back to doing what he does, maybe for a year or two at a time? The learning curve at least ought to be minimal, for a RedHat refugee ...
You'd still need to check what updates are installed and fix problems, but it is a good, over-simplified summary.
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
Dear P.,
Yes, this has happened EVERY time I have installed CentOS 4.0 (the only version I have ever installed). I also boot up into rescue mode and then issue just one grub command,
grub-install /dev/i2o/hda (or whatever your device is called)
This fixes the problem permanently.
Stephen Westrip MetaFour
Peter Farrow wrote:
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Wed, 2005-06-29 at 14:02 +0100, Stephen Westrip wrote:
Dear P.,
Yes, this has happened EVERY time I have installed CentOS 4.0 (the only version I have ever installed). I also boot up into rescue mode and then issue just one grub command,
grub-install /dev/i2o/hda (or whatever your device is called)
This fixes the problem permanently.
Stephen Westrip MetaFour
Peter Farrow wrote:
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
Just so you don't think this is a CentOS only issue: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149587
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149064
Also ...
Grub does not write to the mirror drive ... which can cause issues. It is recommended that you do a grub install (or Peter's solution) on both drives on your first reboot after a software mirror install. (Or put something in your ks file to do it post isntall for kickstart).
On Wed, 2005-06-29 at 08:26 -0500, Johnny Hughes wrote:
On Wed, 2005-06-29 at 14:02 +0100, Stephen Westrip wrote:
Dear P.,
Yes, this has happened EVERY time I have installed CentOS 4.0 (the only version I have ever installed). I also boot up into rescue mode and then issue just one grub command,
grub-install /dev/i2o/hda (or whatever your device is called)
This fixes the problem permanently.
Stephen Westrip MetaFour
Peter Farrow wrote:
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
Just so you don't think this is a CentOS only issue: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149587
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149064
Also ...
Grub does not write to the mirror drive ... which can cause issues. It is recommended that you do a grub install (or Peter's solution) on both drives on your first reboot after a software mirror install. (Or put something in your ks file to do it post isntall for kickstart).
Also ... this isn't always a problem on all hardware during a normal boot, but it is always a problem on any hardware if you boot from the raid mirror drive. So fixing it prior to placing the Raid machine into service is a good thing.
On Wed, 2005-06-29 at 08:30 -0500, Johnny Hughes wrote:
On Wed, 2005-06-29 at 08:26 -0500, Johnny Hughes wrote:
On Wed, 2005-06-29 at 14:02 +0100, Stephen Westrip wrote:
Dear P.,
Yes, this has happened EVERY time I have installed CentOS 4.0 (the only version I have ever installed). I also boot up into rescue mode and then issue just one grub command,
grub-install /dev/i2o/hda (or whatever your device is called)
This fixes the problem permanently.
Stephen Westrip MetaFour
Peter Farrow wrote:
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
Just so you don't think this is a CentOS only issue: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149587
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149064
Also ...
Grub does not write to the mirror drive ... which can cause issues. It is recommended that you do a grub install (or Peter's solution) on both drives on your first reboot after a software mirror install. (Or put something in your ks file to do it post isntall for kickstart).
Also ... this isn't always a problem on all hardware during a normal boot, but it is always a problem on any hardware if you boot from the raid mirror drive. So fixing it prior to placing the Raid machine into service is a good thing.
And one last thing :) ...
I normally do not software raid my boot partition ... I will create a separate partition on both drives, do grub install, and mount one as /boot2 ... then write a script to keep the partition rsynced with the main /boot partition.
That is just how I do it ... and it works for Raid1 or Raid5 :)
Quoting Johnny Hughes mailing-lists@hughesjr.com:
That is just how I do it ... and it works for Raid1 or Raid5 :)
Are you saying your /boot2 is RAID 5? You will not be able to boot of it, in case disk with /boot fails. Whatever partition contains kernel and initrd images, must be either normal partition or two component RAID 1 (2+ component RAID 1, usually called RAID 0+1 and RAID 1+0 will not work). In case it is RAID 1, it works since each component of RAID 1 can be accessed as "normal" partition by boot loader (since boot loader only reads, and do not write, this is OK). For any other RAID configuration, components can not be accessed as "normal" partitions, the data on the volume is accessible as RAID only, and boot loaders are not capable of doing something like that.
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
On Wed, 2005-06-29 at 09:59 -0500, alex@milivojevic.org wrote:
Quoting Johnny Hughes mailing-lists@hughesjr.com:
That is just how I do it ... and it works for Raid1 or Raid5 :)
Are you saying your /boot2 is RAID 5? You will not be able to boot of it, in case disk with /boot fails. Whatever partition contains kernel and initrd images, must be either normal partition or two component RAID 1 (2+ component RAID 1, usually called RAID 0+1 and RAID 1+0 will not work). In case it is RAID 1, it works since each component of RAID 1 can be accessed as "normal" partition by boot loader (since boot loader only reads, and do not write, this is OK). For any other RAID configuration, components can not be accessed as "normal" partitions, the data on the volume is accessible as RAID only, and boot loaders are not capable of doing something like that.
No ... /boot2 is not on raid at all ... it is just a non-raid partition.
I normally build machines (that I am going to do RAID on) with X number of the same hard drives. (Lets say {4} 80gb drives as an example)
so ... I would have a 100mb partition on each drive that is not raid ... on the first drive it is mounted as /boot ... on the others it is /boot2, /boot3, /boot4 ... etc. Rsync boot to the others (and make the partitions bootable)
so ... boot is backed up ... everything else is RAID5
Hi Johnny,
would you share that script with me possibly?
P.
Johnny Hughes wrote:
On Wed, 2005-06-29 at 08:30 -0500, Johnny Hughes wrote:
On Wed, 2005-06-29 at 08:26 -0500, Johnny Hughes wrote:
On Wed, 2005-06-29 at 14:02 +0100, Stephen Westrip wrote:
Dear P.,
Yes, this has happened EVERY time I have installed CentOS 4.0 (the only version I have ever installed). I also boot up into rescue mode and then issue just one grub command,
grub-install /dev/i2o/hda (or whatever your device is called)
This fixes the problem permanently.
Stephen Westrip MetaFour
Peter Farrow wrote:
Dear All,
I have recently had a bad experiences with Centos 4.0 and 4.1 when installing onto mirrored IDE drives using software mirror.
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
P.
Just so you don't think this is a CentOS only issue: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149587
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149064
Also ...
Grub does not write to the mirror drive ... which can cause issues. It is recommended that you do a grub install (or Peter's solution) on both drives on your first reboot after a software mirror install. (Or put something in your ks file to do it post isntall for kickstart).
Also ... this isn't always a problem on all hardware during a normal boot, but it is always a problem on any hardware if you boot from the raid mirror drive. So fixing it prior to placing the Raid machine into service is a good thing.
And one last thing :) ...
I normally do not software raid my boot partition ... I will create a separate partition on both drives, do grub install, and mount one as /boot2 ... then write a script to keep the partition rsynced with the main /boot partition.
That is just how I do it ... and it works for Raid1 or Raid5 :)
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Quoting Johnny Hughes mailing-lists@hughesjr.com:
Grub does not write to the mirror drive ... which can cause issues. It is recommended that you do a grub install (or Peter's solution) on both drives on your first reboot after a software mirror install. (Or put something in your ks file to do it post isntall for kickstart).
Provided /boot is (hd0,0) and (hd1,0). Edit if different, or if /boot is not separate partition. My mail reader is probably going to wrap around install commands. The install commands should be one long line each. So you should have three lines between "grub" and "EOF": "install", "install", "quit". Join the lines if displayed differently.
echo "Fixing GRUB (check /var/log/grubfix.log for errors)" grep -v 'root (' /boot/grub/grub.conf > /boot/grub/newgrub.conf mv -f /boot/grub/newgrub.conf /boot/grub/grub.conf grub --batch <<EOF >/var/log/grubfix.log 2>&1 install --stage2=/boot/grub/stage2 (hd0,0)/grub/stage1 (hd0) (hd0,0)/grub/stage2 p (hd0,0)/grub/grub.conf install --stage2=/boot/grub/stage2 (hd1,0)/grub/stage1 (hd1) (hd1,0)/grub/stage2 p (hd1,0)/grub/grub.conf quit EOF
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
you've obviously not been doing grub that many years :)
grub install on software mirror arrays is finally properly done in FC4 and therefore the future RHEL5.
No you're quite right,
I have found grub to be poorly documented awkward to use and buggy,
LILO is much better, never had this nonsense with LILO :-)
But Lilo is not an option anymore on the installer... :-(
P.
Feizhou wrote:
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
I've been doing Redhat for many years and never seen this problem before......anybody else seen this
you've obviously not been doing grub that many years :)
grub install on software mirror arrays is finally properly done in FC4 and therefore the future RHEL5. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Peter Farrow wrote:
No you're quite right,
I have found grub to be poorly documented awkward to use and buggy,
Buggy? That I don't buy...I've seen my fair share of LI :)
Awkward to use? You must be joking right?
LILO is much better, never had this nonsense with LILO :-)
and also supports installing on LVM2 although there is a patch for grub out there.
But Lilo is not an option anymore on the installer... :-(
yeah...but I don't miss it after I learnt the power of grub.
To get over the install on mirrored / and for other reasons, I have gone pxegrub. I also use pxegrub to bring back online boxes that have encountered lilo/grub loading problems.
Nothing beats grub's ability to let you roam around the filesystem to find out just what kernels were installed and what their configuration parameters were and then load a kernel and its parameters on the spot.
Quoting Feizhou feizhou@graffiti.net:
Nothing beats grub's ability to let you roam around the filesystem to find out just what kernels were installed and what their configuration parameters were and then load a kernel and its parameters on the spot.
I agree that Grub is more featurefull boot loader with many nice options. However, for my servers I simply want them to boot without human intervention. Grub fails to do that more often than LILO. Grub needs extra work to get it right on my servers (that have mirrored disks). I don't care about nice graphical menu (there's nobody in server room to watch it). I don't care about Grub's CLI (there's nobody in server room in the middle of the night to use it). I just want system to boot every time.
For dedicated Linux servers, LILO works perfectly, out of the box, no additional work required. Those boxes are simple to boot, no fancy stuff in boot loader config files. Grub is better when you have "complicated boot configuraion" box. And "complicated boot configuration" boxes are usually desktops and laptops, especially those in hands of developers.
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
alex@milivojevic.org wrote:
Quoting Feizhou feizhou@graffiti.net:
Nothing beats grub's ability to let you roam around the filesystem to find out just what kernels were installed and what their configuration parameters were and then load a kernel and its parameters on the spot.
I agree that Grub is more featurefull boot loader with many nice options. However, for my servers I simply want them to boot without human intervention. Grub fails to do that more often than LILO. Grub needs extra work to get it right on my servers (that have mirrored disks). I don't care about nice graphical menu (there's nobody in server room to watch it). I don't care about Grub's CLI (there's nobody in server room in the middle of the night to use it). I just want system to boot every time.
Ha. Put multiple dhcp servers and tftp servers and the only thing that will prevent pxegrub from booting your box will be a network problem where the issue of whether box is up or not becomes moot. pxegrub will overcome any bootloader issues unless of course you have faulty RAM/NIC.
For dedicated Linux servers, LILO works perfectly, out of the box, no additional work required. Those boxes are simple to boot, no fancy stuff in boot loader config files. Grub is better when you have "complicated boot configuraion" box. And "complicated boot configuration" boxes are usually desktops and laptops, especially those in hands of developers.
I completely disagree with you here since I use grub over pxe to install/boot my servers and avoid any local bootloading problems.
In one sentence, grub is poorly documented, over complicated pants, and lilo ROCKS...
Feizhou wrote:
alex@milivojevic.org wrote:
Quoting Feizhou feizhou@graffiti.net:
Nothing beats grub's ability to let you roam around the filesystem to find out just what kernels were installed and what their configuration parameters were and then load a kernel and its parameters on the spot.
I agree that Grub is more featurefull boot loader with many nice options. However, for my servers I simply want them to boot without human intervention. Grub fails to do that more often than LILO. Grub needs extra work to get it right on my servers (that have mirrored disks). I don't care about nice graphical menu (there's nobody in server room to watch it). I don't care about Grub's CLI (there's nobody in server room in the middle of the night to use it). I just want system to boot every time.
Ha. Put multiple dhcp servers and tftp servers and the only thing that will prevent pxegrub from booting your box will be a network problem where the issue of whether box is up or not becomes moot. pxegrub will overcome any bootloader issues unless of course you have faulty RAM/NIC.
For dedicated Linux servers, LILO works perfectly, out of the box, no additional work required. Those boxes are simple to boot, no fancy stuff in boot loader config files. Grub is better when you have "complicated boot configuraion" box. And "complicated boot configuration" boxes are usually desktops and laptops, especially those in hands of developers.
I completely disagree with you here since I use grub over pxe to install/boot my servers and avoid any local bootloading problems. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Peter Farrow wrote:
In one sentence, grub is poorly documented, over complicated pants, and lilo ROCKS...
Yeah lilo rocks when it works.
but you cannot create a lilo floppy for a server whose kernel you have no idea about nor does it give you a CLI that will allow you to find and load a kernel from the box.
poorly documented? over complicated?
I beg to differ. Commands are so simple.
1) Switch to filesystem at this location and make it root. root (device:part) you can also use tab to do command completion so typing root ([tab] will then give you the list of devices available and their part. 2) another command for loading a kernel image with parameters kernel /path/to/image parameters kernel (device:part)/path/to/image parameters 3) another command for loading an initrd image initrd /path/to/initrdimage initrd (device:path)/path/to/initrdimage
configuration for the boot menu is basically a grouping of the above commands.
And for 2) and 3) you can also use tab for command completion to find the kernel/initrd image files.
lilo is great when it works otherwise it is just a limited bootloader.
Quoting Peter Farrow peter@farrows.org:
But Lilo is not an option anymore on the installer... :-(
It's still an option if you use kickstart. Simply place this line into your kickstart file:
bootloader --location=mbr --useLilo
You might also want to place "-grub" line into %packages section to prevent grub package from being installed.
I haven't done interactive installation for a long time (I'm mostly doing kickstart installations), but I do remember somebody writing that LILO option is hidden under "Advanced" button on intallation screen. I don't know if that is correct.
Sounds like you are doing CentOS/Fedora/RH installations regulary? You should consider doing it using kickstart, it's going to save you lot of time (no more clicking to select same options time and time again). You just boot of CD, type "linux ks=location_of_ks.cfg" on boot prompt, go for a coffe, when you are back, the system is installed. If you are familiar with PXE, you can even have the system boot from the network and install itself automatically.
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
Quoting Peter Farrow peter@farrows.org:
The install goes through ok, but when the system comes to reboot, it seems the boot loader fails. I've seen fail on the first reboot or after a random number of reboots, across different hardware as wide and varied as you can get, the common factor is software raid, (mirrored drives).
It either fails just displaying "GRUB" or fails with a flashing cursor at the bootup time.
Booting a rescue cd, doing a chroot /mnt/sysimage and then the following grub commands fixes the issue permanently, so that it never fails again,
grub
grub-> root (hd0,0) grub-> setup (hd0) grup-> root (hd1,0) grub-> setup (hd1) grub-> quit
it seems to me the Anaconda installer doesn't do something quite right which leads to this random style failure. Its only random as to when it will do it, unless you do this, it will do it at some point.....
Yup, happened to me. With a twist. It seems in my case it is not repairable.
I have 4 machines at home (3 desktops, one laptop). Desktops were fine, laptop shows the symptoms you described. The only thing, the root/setup as described above (or alternative grub-install) does not fix the problem. The laptop is rather standard install, three physical partitions (WinXP, /boot, LVM). WinXP and /boot under 8GB (max addressable by the laptop's BIOS). I can manually type root, kernel and initrd commands on the Grub prompt (the way they appear in grub.conf file), and the system will boot just fine. Even the file name completition works when I hit tab key. So Grub can definitely access the /boot partition. But for whatever reason it fails to read grub.conf and/or display menu screen. It's kind of annoying having to manually type several commands each time I want to use the laptop, but I got used to it.
One more detail. I had Fedora Core 3 installed on the laptop prior to CentOS 4. Same disk configuration, same partition sizes. And Grub worked out of the box as it should.
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
On Thu, 2005-23-06 at 15:58 +0300, Sukru TIKVES wrote:
This is a high traffic mailing list, thus it's very normal to get many emails in a days.
The CentOS list is low volume, in my humble opinion. The Fedora Core user list is a lot busier. But, the highest volume list I've ever subscribed to is the Asterisk list. That list is insane...lots of info though!
Regards,
Ranbir
"High" and "Low" are relative. And I guess it was "high" for him. Anyways it does not matter that much :)
-- sukru
The CentOS list is low volume, in my humble opinion. The Fedora Core user list is a lot busier. But, the highest volume list I've ever subscribed to is the Asterisk list. That list is insane...lots of info though!
Regards,
Ranbir
On Thu, 23 Jun 2005, Sukru TIKVES wrote:
This is a high traffic mailing list, thus it's very normal to get many emails in a days.
That's at least equally obvious on gmane.
It's easy to unsubscribe.
I'm not trying to! I wrote :
I went to the web site for my subscription and tried to disable delivery;
-- it's not the same thing at all.
Most lists -- including this one, judging from one of the inappropriate denials I got -- require you to be subscribed before you can post, including posting by gmane; and it's a good thing they do, since that diminishes the spam.
All lists running on Mailman *also* allow you to be subscribed but not receiving email -- if their instance of Mailman is running properly. This one's apparently is not.
beartooth wrote:
I subscribed, and am getting emails hand over fist. They'll bust my inbox quota if I miss a day.
I went to the web site for my subscription and tried to disable delivery; I'll follow this list on gmane or not at all -- and that works with every other mailman list I've tried.
It wouldn't let me -- kept reverting to the password screen; I kept giving it the password, and making the change; and next time, the change was not there -- even though my password was letting me in.
Weird, I just tried it and it worked. Are you behind a proxy?
I sent an email to the address specified on the subscription website, asking for help. It sent me a piece of boilerplate in reply telling me I can't post (wanna bet?) -- because I'm not subscribed!
On Thu, 23 Jun 2005 13:25:24 -0400, Ugo Bellavance wrote:
Weird, I just tried it and it worked. Are you behind a proxy?
Hmmm.... Yes, Privoxy. But it doesn't impair my handling other Mailman subscriptions. (It's on my private machine, btw, not a corporate one; I'm an old retired fart, with no corporate anything any more. Ah, bliss ...)
To answer someone else's question, I also run half a dozen browsers, with different settings; it's possible I tried with one refusing cookies -- but iirc, I tried three browsers, one of them almost certainly Opera 8.01. Could the site be one that thinks Opera doesn't do cookies even though it does? There are some ...
Anyway, one of the moderators very kindly changed the setting for me; so I'm a happy camper now.
beartooth wrote:
I went to the web site for my subscription and tried to disable delivery; I'll follow this list on gmane or not at all -- and that works with every other mailman list I've tried.
Great idea! I went ahead and did it too just now. Thanks.
It wouldn't let me -- kept reverting to the password screen; I kept giving it the password, and making the change; and next time, the change was not there -- even though my password was letting me in.
I didn't have a problem disabling delivery at all. Do you have cookies enabled?
johnn
-
alex＠milivojevic.org -
Alexander Dalloz -
Bear Tooth -
beartooth
-
Feizhou -
Johnn Tan -
Johnny Hughes -
Kanwar Ranbir Sandhu -
Lamar Owen -
Peter Farrow -
Rodrigo Barbosa -
Stephen Westrip -
Steve Huff -
Sukru TIKVES -
Ugo Bellavance