Re: testing upgrade = dog's breakfast
On Tue, 2006-05-09 at 22:24 +0200, lee wrote: > On Tue, May 09, 2006 at 11:35:42AM -0400, Greg Folkert wrote: > > > Well, if you had installed Sarge or Etch, they would have upgraded your > > Exim v3 install to Exim v4. I've done the transition a couple of hundred > > times, it has never failed me except on a hand crafted config. > > Neither Sarge, nor testing upgrade Exim3 to Exim4. Default on Sarge > was Exim3 when I installed it, default on testing is Exim4. Since the > configuration is much different between the two versions, it is _very > good_ policy that Exim3 is not upgraded unless you do it yourselfe. Okay, where were you when it asks: "or do you have a handcrafted exim3 config" and do the change over manually (unconfigured until you fix it) Every single machine I have upgraded is now part of a cluster of machines that use a standard "little files" setup and I just have a package I install that puts everything in there I need. (along with my vhosting setup) > A switch from lilo to grub is also not made in the upgrading process, > but that's something I would whish to have an automatic upgrade for > --- though that's probably difficult to do unless you stick to the > kernel packages from the distribution instead of making your own > kernel. Again, if Lilo *IS* installed it does not change to grub. If grub *IS* installed it does not go BACK to Lilo unless you changed your "/" filesystem to XFS. > You likewise get Cyrus 1.5 with Sarge. I don't know what it upgrades > to when upgrading to testing, but imho it _should_ be upgraded to a > more recent 1.5 only. Cyrus 2.x is much different from 1.5, and an > upgrade is something that is definitely to be done manually, after > reading the docs and, if possible, after having the upgrading tried > out in a testing environment on a copy of the real data. Please read what I write and not assume what you think I wrote. -- greg, [EMAIL PROTECTED] The technology that is Stronger, Better, Faster: Linux Use Debian GNU/Linux, its a bazaar thing NOTICE: Due to Presidential Executive Orders, the National Security Agency may have read this email without warning, warrant, or notice, and certainly without probable cause. They may do this without any judicial or legislative oversight. You have no recourse nor protection. signature.asc Description: This is a digitally signed message part
Re: testing upgrade = dog's breakfast
At 1147355422 past the epoch, Georgi Alexandrov wrote: > Just complaining and stating over and over again that > something doesn't work when you are asked specific > questions about the way you do it is childish, really! The first message to this thread was fairly ranty, but his follow-ups have been informative. If you don't like the wording or attitude of a poster, just don't reply to them. -- Jon Dowland http://alcopop.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
Hans Vogelsberger wrote: > Georgi Alexandrov schrieb: > >> This guy was asked to explain . . . distrib. Yet he still just cries >> like baby. > >> Well, cry off the list please! >> > Are there no lessons for the most primitive forms of good manners in > the country you live? > > Hans > > Yes, but they probably have limits. Just complaining and stating over and over again that something doesn't work when you are asked specific questions about the way you do it is childish, really! have a nice day, -- regards, Georgi Alexandrov key server - http://pgp.mit.edu/ :: key id - 0x37B4B3EE key fingerprint - E429 BF93 FA67 44E9 B7D4 F89E F990 01C1 37B4 B3EE signature.asc Description: OpenPGP digital signature
Re: testing upgrade = dog's breakfast
Georgi Alexandrov schrieb: This guy was asked to explain . . . distrib. Yet he still just cries like baby. Well, cry off the list please! Are there no lessons for the most primitive forms of good manners in the country you live? Hans -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
lee wrote: > Maybe the upgrade path from stable to testing has not been settled > yet. You may get better results with installing testing rather than > doing an upgrade. > > Actually, my point of view is that you can be better off with testing > when setting up a system that has to be kept up to date for years to > come, one reason for that being the hassle of eventually sticking with > older versions of software after an upgrade. For the last three > servers I set up at work, I decided to use testing, and they work > flawlessly. I'm also running it at home. > > > GH > The path have been settled well, I've done sarge > etch many times with no problem. This guy was asked to explain his upgrade procedure and show his apt configs as 2.4.18 really is the default 2.4 kernel with the old woody distrib. Yet he still just cries like baby. Well, cry off the list please! ;-) -- regards, Georgi Alexandrov key server - http://pgp.mit.edu/ :: key id - 0x37B4B3EE key fingerprint - E429 BF93 FA67 44E9 B7D4 F89E F990 01C1 37B4 B3EE signature.asc Description: OpenPGP digital signature
Re: testing upgrade = dog's breakfast
>>"Jon Dowland" <[EMAIL PROTECTED]> wrote in message >>news:[EMAIL PROTECTED] >> >> >>>With a stable install (sarge) lilo should not be installed in the first >>>place, let alone upgraded. >>> Why not? Lilo is more capable than grub with respect to partition formats and afaik, is the only way to achieve a system that entirely uses LVM. For all other typical uses on x86 systems, I do still prefer grub. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
On Tue, May 09, 2006 at 02:38:23PM -0600, Art Edwards wrote: > My current plan is to rebuild a stable box and leave testing to the > hobbyists. Maybe the upgrade path from stable to testing has not been settled yet. You may get better results with installing testing rather than doing an upgrade. Actually, my point of view is that you can be better off with testing when setting up a system that has to be kept up to date for years to come, one reason for that being the hassle of eventually sticking with older versions of software after an upgrade. For the last three servers I set up at work, I decided to use testing, and they work flawlessly. I'm also running it at home. GH -- Life is hard; and then you die. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
I probably was not clear about what happened. Prior to switching to testing, I was running Grub. Also, the frozen messages happened prior to the switch. It seems that exim3 and mutt did not play well together (thunderbird worked well, however.) I was also running the 2.6.12 kernel as default under grub in stable. The mouse worked perfectly with the 2.6.12 cursor in the stable distribution. After the upgrade, lilo was the default loader (not my choice), and the 2.4.18b kernel was the only kernel built into lilo. This latter kernel result was probably because lilo simply built a Linux stanza. The mouse worked perfectly under the 2.4.18b kernel. However, when I added the stanza for the 2.6.12 kernel and booted, the mouse was completely erractic. On a console I received messages about losing synchronization. My current plan is to rebuild a stable box and leave testing to the hobbyists. Art Edwards On Tue, May 09, 2006 at 10:24:59PM +0200, lee wrote: > On Tue, May 09, 2006 at 11:35:42AM -0400, Greg Folkert wrote: > > > Well, if you had installed Sarge or Etch, they would have upgraded your > > Exim v3 install to Exim v4. I've done the transition a couple of hundred > > times, it has never failed me except on a hand crafted config. > > Neither Sarge, nor testing upgrade Exim3 to Exim4. Default on Sarge > was Exim3 when I installed it, default on testing is Exim4. Since the > configuration is much different between the two versions, it is _very > good_ policy that Exim3 is not upgraded unless you do it yourselfe. > > A switch from lilo to grub is also not made in the upgrading process, > but that's something I would whish to have an automatic upgrade for > --- though that's probably difficult to do unless you stick to the > kernel packages from the distribution instead of making your own > kernel. > > You likewise get Cyrus 1.5 with Sarge. I don't know what it upgrades > to when upgrading to testing, but imho it _should_ be upgraded to a > more recent 1.5 only. Cyrus 2.x is much different from 1.5, and an > upgrade is something that is definitely to be done manually, after > reading the docs and, if possible, after having the upgrading tried > out in a testing environment on a copy of the real data. > > > GH > -- > Life is hard; and then you die. > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
On Tue, May 09, 2006 at 11:35:42AM -0400, Greg Folkert wrote: > Well, if you had installed Sarge or Etch, they would have upgraded your > Exim v3 install to Exim v4. I've done the transition a couple of hundred > times, it has never failed me except on a hand crafted config. Neither Sarge, nor testing upgrade Exim3 to Exim4. Default on Sarge was Exim3 when I installed it, default on testing is Exim4. Since the configuration is much different between the two versions, it is _very good_ policy that Exim3 is not upgraded unless you do it yourselfe. A switch from lilo to grub is also not made in the upgrading process, but that's something I would whish to have an automatic upgrade for --- though that's probably difficult to do unless you stick to the kernel packages from the distribution instead of making your own kernel. You likewise get Cyrus 1.5 with Sarge. I don't know what it upgrades to when upgrading to testing, but imho it _should_ be upgraded to a more recent 1.5 only. Cyrus 2.x is much different from 1.5, and an upgrade is something that is definitely to be done manually, after reading the docs and, if possible, after having the upgrading tried out in a testing environment on a copy of the real data. GH -- Life is hard; and then you die. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
Thanks for at least looking at a fairly irate posting. Actually, I installed stable in February using the netinstall sarge disk. Art Edwards On Tue, May 09, 2006 at 12:43:26PM -0400, Joe Smith wrote: > > "Jon Dowland" <[EMAIL PROTECTED]> wrote in message > news:[EMAIL PROTECTED] > >At 1147164438 past the epoch, Art Edwards wrote: > >>On Friday night I updated a stable box to a testing box. Afterwards: > > > >What upgrade method did you use? > > > >>1. Lilo was promoted to the default boot loader > > > >With a stable install (sarge) lilo should not be installed in the first > >place, let alone upgraded. > > > >>I should also point out that stable installed exim, instead of exim4 > >>so I had forzen messages for 8 weeks under mutt. This is under stable. > > > >Sarge (stable) defaults to exim4 for your MTA. > > > >These two things lead me to believe you have some kind-of hybrid or > >partially updated woody system rather than sarge -> etch. > > I think he may have accidentally skipped sarge entirely, trying a > woody->etch transition. > That may explain some of these problems. > > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact > [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
"Jon Dowland" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] At 1147164438 past the epoch, Art Edwards wrote: On Friday night I updated a stable box to a testing box. Afterwards: What upgrade method did you use? 1. Lilo was promoted to the default boot loader With a stable install (sarge) lilo should not be installed in the first place, let alone upgraded. I should also point out that stable installed exim, instead of exim4 so I had forzen messages for 8 weeks under mutt. This is under stable. Sarge (stable) defaults to exim4 for your MTA. These two things lead me to believe you have some kind-of hybrid or partially updated woody system rather than sarge -> etch. I think he may have accidentally skipped sarge entirely, trying a woody->etch transition. That may explain some of these problems. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: testing upgrade = dog's breakfast
On Tue, 2006-05-09 at 08:47 -0600, Art Edwards wrote: > On Friday night I updated a stable box to a testing box. Afterwards: > > 1. Lilo was promoted to the default boot loader First off, are you using XFS as your root filesystem? If so, there is a problem with grub detecting things properly and it freezes the XFS improperly when installing itself. Therefore LILO would be the default booter on that. > 2. I was running a 2.4.18 kernel (instead of the 2.6.12 kernel under > stable) Well, that tells me you at one time were running Woody, or are now... a kind of pseudo downgrade may have happened. What does your "/etc/apt/source.list" look like? > 3. Gnome panels don't work, so I'm stuck in KDE Well, I had a similar problem when going from GNOME 1.4 to GNOME 2.0 (and 2.2 and 2.4 and 2.6 and 2.8 and 2.10 and 2.12 and not with 2.14) It took removing the GNOME user config settings to get GNOME back. > I updated the lilo.conf file to include the 2.6.12 kernel. Under that > kernel > > 1. the network doesn't come up. (It does after I use modprobe to load > the appropriate ethernet driver > 2. The mouse seems to have had a nervous breakdown. > 3. When I try to unstall grub, it can't find the stage1file that I > can > using ls. Previous isn't important to your problem... because it is a symptom of a greater problem. > I realize that testing is not stable (that's why there are two > different names), but this is the sort if behavior I would expect from > sid (alpha test?) instead of an obstensibly beta test set. Actually if you really think about it there are Currently 5 active distributions of Debian right now. 1. Old Stable: Woody 2. Stable: Sarge 3. Testing:Etch 4. Unstable: SID (Still In Development) 5. Experimental: Experimental Old Stable is what I think you have back-graded to... maybe not, but something is a miss. We'll need to see your "sources.list" and any "apt.conf" or "preferences" files you may have in "/etc/apt" > I should also point out that stable installed exim, instead of exim4 > so I had forzen messages for 8 weeks under mutt. This is under stable. Well, if you had installed Sarge or Etch, they would have upgraded your Exim v3 install to Exim v4. I've done the transition a couple of hundred times, it has never failed me except on a hand crafted config. Check my other reply to your "grub and lilo" thread. It'll help you get grub back as the default bootloader. -- greg, [EMAIL PROTECTED] The technology that is Stronger, Better, Faster: Linux Use Debian GNU/Linux, its a bazaar thing NOTICE: Due to Presidential Executive Orders, the National Security Agency may have read this email without warning, warrant, or notice, and certainly without probable cause. They may do this without any judicial or legislative oversight. You have no recourse nor protection. signature.asc Description: This is a digitally signed message part
Re: testing upgrade = dog's breakfast
At 1147164438 past the epoch, Art Edwards wrote: > On Friday night I updated a stable box to a testing box. Afterwards: What upgrade method did you use? > 1. Lilo was promoted to the default boot loader With a stable install (sarge) lilo should not be installed in the first place, let alone upgraded. > I should also point out that stable installed exim, instead of exim4 > so I had forzen messages for 8 weeks under mutt. This is under stable. Sarge (stable) defaults to exim4 for your MTA. These two things lead me to believe you have some kind-of hybrid or partially updated woody system rather than sarge -> etch. -- Jon Dowland http://alcopop.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
testing upgrade = dog's breakfast
On Friday night I updated a stable box to a testing box. Afterwards: 1. Lilo was promoted to the default boot loader 2. I was running a 2.4.18 kernel (instead of the 2.6.12 kernel under stable) 3. Gnome panels don't work, so I'm stuck in KDE I updated the lilo.conf file to include the 2.6.12 kernel. Under that kernel 1. the network doesn't come up. (It does after I use modprobe to load the appropriate ethernet driver 2. The mouse seems to have had a nervous breakdown. 3. When I try to unstall grub, it can't find the stage1file that I can using ls. I realize that testing is not stable (that's why there are two different names), but this is the sort if behavior I would expect from sid (alpha test?) instead of an obstensibly beta test set. I should also point out that stable installed exim, instead of exim4 so I had forzen messages for 8 weeks under mutt. This is under stable. Art Edwards -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]