[opensuse-factory] YaST2 System Update not too bright
Prior to updating from about beta2 to current factory tree today, I downloaded and installed the 2.6.18.2-33-default kernel with rpm -ivh so that the previous 2.6.18.2-5 kernel would remain installed. I forgot to reboot prior to running System Update, which proceeded to install 2.6.18.2-33 again, and remove 2.6.18.2-5. :-( -- "Let your conversation be always full of grace." Colossians 4:6 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Felix Miata schrieb:
Prior to updating from about beta2 to current factory tree today, I downloaded and installed the 2.6.18.2-33-default kernel with rpm -ivh so that the previous 2.6.18.2-5 kernel would remain installed. I forgot to reboot prior to running System Update, which proceeded to install 2.6.18.2-33 again, and remove 2.6.18.2-5. :-(
Don't take this as something evil, but I have wondered since quite some time what the purpose of these status report mails to the opensuse-factory list is. If you think that any of the involved software components - be it YaST, the kernel package, the bootloader config scripts - misbehaves, file a bug. That's what Bugzilla is for. If you need help setting up your system for a particular purpose, try to state as clearly as possible what you're trying to achieve. I read the text three times now and it's still not entirely clear to me what this is all about: - If you think that a part of the system misbehaves, file a bug including what has to change: YaST, the kernel, perl-bootloader, something else... And describe the difference between the actual and the desired behaviour in such a way that it's understandable without any knowledge beyond your report. - If you just want to update your system to something that resembles what will be openSUSE 10.2 GM, forget about factory now. There's nothing interesting to see in factory right now. - If you want to have a special setup, try to find out if it's possible at all. Last time I checked, the handling of multiple equally named kernel packages in YaST was just a vision. Don't know if something changed, but right now I'd say YaST doesn't support that. In addition, you might want to check whether the old kernel really disappeared or whether it's just the bootloader entry that was missing. Note that during a not quite correctly done system upgrade, the _old_ version of YaST and its libaries are running. This means that you might suffer from bugs that have been fixed long ago. perl-Bootloader and yast2-bootloader had lots of them. Beta2 is very old. That's why the recommended upgrade way is booting from the installation media and choosing the system upgrade option from there. Doing a system upgrade within the running system means that you'll see and report old, known and already fixed bugs. Andreas --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On 2006/12/02 21:42 (GMT+0100) Andreas Hanke apparently typed:
Felix Miata schrieb:
Prior to updating from about beta2 to current factory tree today, I downloaded and installed the 2.6.18.2-33-default kernel with rpm -ivh so that the previous 2.6.18.2-5 kernel would remain installed. I forgot to reboot prior to running System Update, which proceeded to install 2.6.18.2-33 again, and remove 2.6.18.2-5. :-(
Don't take this as something evil, but I have wondered since quite some time what the purpose of these status report mails to the opensuse-factory list is.
When one doesn't know if observed behavior is or is not a bug, one asks here rather than cluttering bugzilla and its QA people with another likely invalid bug.
If you think that any of the involved software components - be it YaST, the kernel package, the bootloader config scripts - misbehaves, file a bug. That's what Bugzilla is for.
Bugzilla is not for teaching people what program behavior is expected, which is what happens when somebody files a bug without knowing if the observed behavior is that way by design. In this case maybe my forgetting to reboot to engage the new kernel means kernel installation wasn't actually complete before the update process started - I don't know, which is a reason why I came here rather than to bugzilla.
If you need help setting up your system for a particular purpose, try to state as clearly as possible what you're trying to achieve. I read the text three times now and it's still not entirely clear to me what this is all about:
The updater installed software that was already installed, a substantial unnecessary load on mirrors, and waste of my time waiting for the unnecessary download from an already slow mirror.
In addition, you might want to check whether the old kernel really disappeared or whether it's just the bootloader entry that was missing.
I checked. It was removed from /boot and /lib/modules, besides the boot menu. That's how I know.
Note that during a not quite correctly done system upgrade, the _old_ version of YaST and its libaries are running. This means that you might suffer from bugs that have been fixed long ago. perl-Bootloader and yast2-bootloader had lots of them. Beta2 is very old.
Beta2 was actually a guess. I don't think the install was more than 3 weeks old, which, like virtually all my installs, was via ftp. The possibility of already fixed bugs is another reason for coming here rather than filing a bug. The kernel upgrade I did manually required perl-Bootloader be upgraded first, which I did manually as well.
That's why the recommended upgrade way is booting from the installation
I don't care what's recommended. I used a method offered. If it's offered, it needs testing.
media and choosing the system upgrade option from there. Doing a system upgrade within the running system means that you'll see and report old, known and already fixed bugs.
I have factory installed on 5 (or 6?) partitions on various hardware combinations expressly for finding stuff that doesn't install or run as intended. I don't burn CD's except on rare occasions, starting from an installation kernel and and initrd downloaded from the mirror immediately before starting in most cases, so that I don't have to type in manually answers to the network and source questions every time, or spend money wasting non-renewable resources. It's disappointing to find so much reported weeks or months or close to a year ago in bugzilla that didn't get fixed before 10.2 "release" state was achieved. There's one in particular I reported against 10.0 (and reproduced on at least 4 entirely different systems using 10.0, 10.1 & factory) long before 10.1's first beta that remains unfixed now. I've lived with it on my 24/7 box nearly a year now. Likely that box will see 10.0 replaced with Mandriva 2007 soon in order that I can stop being constantly annoyed by it. -- "Let your conversation be always full of grace." Colossians 4:6 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Felix Miata schrieb:
The updater installed software that was already installed, a substantial unnecessary load on mirrors, and waste of my time waiting for the unnecessary download from an already slow mirror.
Now I finally got what it's about: You installed the -33 kernel and YaST downloaded and installed the same kernel again. The only way I can imagine this not being a bug is that the "rpm -i" has been done while YaST was already running. That way YaST won't notice the change because it reads the rpmdb on startup and doesn't lock it. You might want to file a bug about this. Of course you will attach the logfiles.
I don't care what's recommended. I used a method offered. If it's offered, it needs testing.
It might be disabled at any time then ;-) (Already happened to other features where it turned out that offering them unsupported doesn't really work and resulted in more frustration than not offering it at all would)
It's disappointing to find so much reported weeks or months or close to a year ago in bugzilla that didn't get fixed before 10.2 "release" state was achieved. There's one in particular I reported against 10.0 (and reproduced on at least 4 entirely different systems using 10.0, 10.1 & factory) long before 10.1's first beta that remains unfixed now. I've lived with it on my 24/7 box nearly a year now. Likely that box will see 10.0 replaced with Mandriva 2007 soon in order that I can stop being constantly annoyed by it.
A general hint: Posting that to this list without the Bugzilla ID doesn't really bring the community any further. Writing something like "I'll switch to Mandriva" is very demotivating because it basically means that the bug doesn't need to be fixed because the only user annoyed by it doesn't use the product anyway. This is very counter-productive. You might, instead, try to be cooperative, i.e. show that you're still interested in the bug being solved by moving it to the next product and/or trying to fix it yourself and attaching a patch or a pointer into the right direction. E.g. by looking at source packages of other distributions where it works, finding out the difference and telling the assignee about it. Without knowing the Bugzilla ID, I can unfortunately only guess that the engineer who is responsible for your bug has a long list of more important bugs to solve that affect more people and you can't see that list because the engineers responsible for openSUSE are responsible for other products as well which are not public. Might be considered frustrating, but at some point every community member has to learn that there are bugs which will never be fixed. This is common to all open source projects. I just tried to find a 2 years old, unfixed bug in Mandriva's Bugzilla and of course it was a piece of cake finding one. Btw. I know that it's about #141443 because it's the only one you have reported against 10.0 that is still open. In that case it's really unfortunate, but as long as an engineer is unable to reproduce it, it's simply impossible to get it fixed. Andreas --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On 2006/12/02 23:13 (GMT+0100) Andreas Hanke apparently typed:
Felix Miata schrieb:
The updater installed software that was already installed, a substantial unnecessary load on mirrors, and waste of my time waiting for the unnecessary download from an already slow mirror.
Now I finally got what it's about: You installed the -33 kernel and YaST downloaded and installed the same kernel again.
The only way I can imagine this not being a bug is that the "rpm -i" has been done while YaST was already running. That way YaST won't notice the change because it reads the rpmdb on startup and doesn't lock it.
It's possible I may have left YaST open while doing the manual kernel install the first time, but today I tried again without forgetting to reboot before updating.
You might want to file a bug about this. Of course you will attach the logfiles.
https://bugzilla.novell.com/show_bug.cgi?id=225674
It's disappointing to find so much reported weeks or months or close to a year ago in bugzilla that didn't get fixed before 10.2 "release" state was achieved. There's one in particular I reported against 10.0 (and reproduced on at least 4 entirely different systems using 10.0, 10.1 & factory) long before 10.1's first beta that remains unfixed now. I've lived with it on my 24/7 box nearly a year now. Likely that box will see 10.0 replaced with Mandriva 2007 soon in order that I can stop being constantly annoyed by it.
A general hint: Posting that to this list without the Bugzilla ID doesn't really bring the community any further. Writing something like
I already raised the issue on the mailing list and don't believe any more can be expected from the community, nor do I want to look like I'm bashing the assignee in a public forum.
"I'll switch to Mandriva" is very demotivating because it basically
I'll switch means the severity of the annoyance is too high. The two distros are nearly equivalent in meeting my needs. Without this bug and one other I haven't even bothered to report, SUSE would remain my preference. I think two releases without a fix is justification for giving up.
means that the bug doesn't need to be fixed because the only user annoyed by it doesn't use the product anyway. This is very counter-productive.
I'm not the only user observing the problem, as you can see reading the bug and reading the mailing list discussion about it. I imagine there must be users who are not bugzilla participants or mailing list subscribers who see it too, and may not even know what's going on.
You might, instead, try to be cooperative, i.e. show that you're still interested in the bug being solved by moving it to the next product and/or trying to fix it yourself and attaching a patch or a pointer into the right direction. E.g. by looking at source packages of other distributions where it works, finding out the difference and telling the assignee about it.
I'm not a programmer, so I won't be providing any patch. If the bug's assignee wants assistance, he needs to indicate so more than simply stating he is unable to reproduce.
Might be considered frustrating, but at some point every community member has to learn that there are bugs which will never be fixed. This is common to all open source projects. I just tried to find a 2 years old, unfixed bug in Mandriva's Bugzilla and of course it was a piece of cake finding one.
Some bugs one can live with, others not. One can only scratch so many times before drawing blood. That's not healthy.
Btw. I know that it's about #141443 because it's the only one you have reported against 10.0 that is still open. In that case it's really unfortunate, but as long as an engineer is unable to reproduce it, it's simply impossible to get it fixed.
I have reproduced it on about 5 or 6 hardware systems using 10.0 and 10.1 and 10.2, which means 100% of about 10 different installs exhibit the problem. Others have reported reproducing it both in the bug and on the mailing list, so probably this has been reproduced and reported on well over a dozen different systems. I think that's enough to indicate an engineer with sufficient interest and time should be able to reproduce it somehow. I do as much as practical using several ttys. X is not my environment of choice, but necessary for graphics and normal web use. So, this bug hits me constantly. A year is just too much, and soon it will be a year. Switching from SUSE seems to be the only option I have to deal with it. -- "Let your conversation be always full of grace." Colossians 4:6 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2006-12-04 at 00:08 -0500, Felix Miata wrote:
means that the bug doesn't need to be fixed because the only user annoyed by it doesn't use the product anyway. This is very counter-productive.
I'm not the only user observing the problem, as you can see reading the bug and reading the mailing list discussion about it. I imagine there must be users who are not bugzilla participants or mailing list subscribers who see it too, and may not even know what's going on.
Folks, I have to concur. I just had to open bugzilla, find what the bug was about, and reproduced it in under 10 seconds. I can not believe that a SuSE engineer is unable to reproduce it. In fact, I have known of this bug for a very long time, but I didn't bother to report. I'm doing it now. My procedure: - leave firefox with focus - I'm using gnome in 10.1 stock plus security patches. - switch to a text console via ctrl-alt-f1, for instance. - switch back to X via alt-f7 - firefox pops up asking whether I want to enable caret browsing (F7 turn caret browsing on (screen shot available)). M f: - I can add more to this. If I switch to another workspace in gnome and back later to an xterm and some other apps, using the keyboard, extra meaningless text is inputted to the application (escape codes, I think). The usuall result in Pine is that it tries to activate a command, usually "unknown command", some other times pulling me out from the current view to the main menu, or trying to lock the app. It is a real nuisance. In midnight commander I get text in its command line (I use both apps a lot). It is random - Do you see that "M f:" before this paragraph? That's what I mean. This has been going for years, in several SuSE versions. I'm sure that if you ask around, many more people will tell you they can reproduce this behaviour. In fact, it was commented a week or two ago a problem with stuck keys in the keyboard, maybe related. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFdCT9tTMYHG2NR9URAtOWAJ9gs/XbkRt3MNQdTWNc89btRT5HwACePd7u vU2H/7A7ALOlPL8jKXpMBnE= =unhi -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Carlos E. R. schrieb:
Folks, I have to concur.
I just had to open bugzilla, find what the bug was about, and reproduced it in under 10 seconds. I can not believe that a SuSE engineer is unable to reproduce it. In fact, I have known of this bug for a very long time, but I didn't bother to report. I'm doing it now.
Strange. I didn't even bother trying to reproduce it because I never noticed it and IMHO initiating a discussion on the mailinglist about switching distros and how disappointing the entire distribution with its 16000 packages is because of *that* is just troublemaking. Now I tried reproducing it and didn't succeed in 5 attempts. Afterwards, I tried again, this time leaving the F7 key pressed a bit longer. That "worked". Guys, press the F7 key shorter. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On 2006/12/04 14:54 (GMT+0100) Andreas Hanke apparently typed:
Now I tried reproducing it and didn't succeed in 5 attempts.
Afterwards, I tried again, this time leaving the F7 key pressed a bit longer. That "worked".
Guys, press the F7 key shorter.
https://bugzilla.novell.com/show_bug.cgi?id=141443#c1 https://bugzilla.novell.com/show_bug.cgi?id=141443#c2 https://bugzilla.novell.com/show_bug.cgi?id=141443#c37 -- "Let your conversation be always full of grace." Colossians 4:6 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2006-12-04 at 14:54 +0100, Andreas Hanke wrote:
Guys, press the F7 key shorter.
That doesn't work for every body, or every time, unfortunately. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFdDuetTMYHG2NR9URAgW5AJ93YHa8PKmA9aWgNq1caAwswPbdZgCfRKQc qitTdmdeSrcUb2AXYcoi6RI= =1wcg -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Felix Miata schrieb:
I do as much as practical using several ttys. X is not my environment of choice, but necessary for graphics and normal web use. So, this bug hits me constantly. A year is just too much, and soon it will be a year. Switching from SUSE seems to be the only option I have to deal with it.
But be sure not to run into this incredible, 2 years old, unfixed bug that kills kittens and drives people nuts: http://qa.mandriva.com/show_bug.cgi?id=5591 Of course it has the mandatory "If ever this bug is not fixed I'll switch to Ubuntu" comment in it. ;-) Andreas --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
participants (3)
-
Andreas Hanke -
Carlos E. R. -
Felix Miata