You are not logged in.

#76 2009-09-04 01:57:43

broch
Banned
From: L.A. California
Registered: 2006-11-13
Posts: 975

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

@shining
if you do not understand my objections regarding kernel26-lts then say that you do not understand instead of calling it bulshit:
according to the maintainer kernel26-lts is going to be based on 2.6.27 (when I asked why not 2.4 I was joking: this is sad that such thing must be explained.)

2.6.27 does not support a lot of newer options, backporting them often will be problematic. What I suggested is to allow system to keep previous kernel version when instaling next kernel. This way system always will have one bootable kernel and this is much easier to mantain than backport ext4, and whatever fs or device driver that may affect system stability.

do you even know what fallback mean ?

ActualIy doubt that you know what fallback is: same stuff as the latest kernel with everything enabled: search forum for complains that falback does not work either. This is not Arch fault obviously. If kernel introduses a bug, the it does not makker if everything is nabled or not.
If you do not understand a difference between last good and fallback, then no point of discussing this.

Last time I installed Arch was more than two years ago, I don't have problems with kernels or system stability in general.

You have no idea what && is, do you ?

don't get too excited.
first linux that I installed was RedHat 4.2, first UNIX was FreeBSD 2.1
First time I used computer (IBM) was in 76' so please don't lecture me about kernels or &&.

You are too sensitive to the critique it seems. So I will leave to to contempalte with awe the above wiki entry.

Last edited by broch (2009-09-04 01:59:02)

Offline

#77 2009-09-04 02:26:38

Allan
Pacman
From: Brisbane, AU
Registered: 2007-06-09
Posts: 11,650
Website

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

broch wrote:

2.6.27 does not support a lot of newer options, backporting them often will be problematic.

And that is why the kernel devs are doing all the back porting upstream.  About the only patch added is to rename ext4dev to ext4 so this backported feature was useful.

broch wrote:

What I suggested is to allow system to keep previous kernel version when instaling next kernel.

Actually, you suggest that pacman had an option to do this for you, which is different and what shining objected to.

broch wrote:

Last time I installed Arch was more than two years ago, I don't have problems with kernels or system stability in general.

Then you have no need for this wiki page....  Move along.

broch wrote:

First time I used computer (IBM) was in 76' so please don't lecture me about kernels or &&.

Um... so what...   you can still be wrong.

broch wrote:

You are too sensitive to the critique it seems.

Hmm....

Offline

#78 2009-09-04 15:06:31

broch
Banned
From: L.A. California
Registered: 2006-11-13
Posts: 975

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

broch wrote:

    You are too sensitive to the critique it seems.

Hmm....

shining in his response used words: "bulshit" (edited and changed to BS already), "stupid", "FUD" each time when he did not agree (all the time in fact) whith what I said. Fix his post in full before doubting what I said.

broch wrote:

    First time I used computer (IBM) was in 76' so please don't lecture me about kernels or &&.

Um... so what...   you can still be wrong.

asking politely not to lecture has nothing to do with being wrong or not. I refered to the simple fact that he mostly even not tried to read what my post before responding. Not sure where did you get the impression about wrong or right from my post. Calling something "bulshit" or "stupid" is insulting.

If he does not agree with my opinion that is fine, I pointed out that I am probably older than him (using computers since 76') and because I did not use insults so he should not either.

When I suggested that keeping backup files is not reasonable, the ansswer was:

Or just buy a bigger disk.

is this advice doing to be added to wiki?

or second option:

Then what does -Sc do ?

the answer would be:
nothing good, as quite often users complain about disk being full. Option left is livecd and cleaning up pacman db. The reason may be not only too small disk but (even if disk is big) too small /var partition. This is an advantage of real backup.


And that is why the kernel devs are doing all the back porting upstream.  About the only patch added is to rename ext4dev to ext4 so this backported feature was useful.

Because kernel version was taken arbitrary (there is no statistical proof that this particular version is better than any stable version before or after) the end effect is going to be the same as with any new Arch kernel/fallback, some users will have problems booting this. Statistically, number of failed boot events will not change unless you suggest that somehow arch devs neglect current kernel and pay attention extra attention to kernel26-lst. If one keeps last good kernel, such problem (failed boot) has much lower chance to happen as this kernel is naturally tested with specific setup. whether one will use pacman or another script or do this manually does not really matter

broch wrote:

    Last time I installed Arch was more than two years ago, I don't have problems with kernels or system stability in general.

Then you have no need for this wiki page....  Move along.

I thought that this is a discussion about methods of keeping system stable. Or not? You are contradictiong original post:

Over in the 'Newbie Corner' webforum, a fairly active thread emerged, entitled, Does anyone really have a "rock stable" Arch?

In response to this thread, I have put together some material for a new wikipage, Enhancing Arch Linux Stability.

Right now, the tips are fairly basic, but they have worked well to ensure that my own system is stable and reliable. No doubt though, this wikipage would benefit from additional community input. Please share your stability tips by updating the wiki, so that everyone might enjoy a stable Arch experience. Thank you.

Last edited by broch (2009-09-04 15:07:04)

Offline

#79 2009-09-04 15:22:06

DonVla
Member
From: Bonn, Germany
Registered: 2007-06-07
Posts: 997

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

... this one is really getting out of hand...

Offline

#80 2009-09-04 16:12:03

zodmaner
Member
Registered: 2007-07-11
Posts: 653

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

DonVla wrote:

... this one is really getting out of hand...

Indeed. But then again forum threads that have the words "Arch" and "stability" in them have always tend to be volatile...

I would suggest that if anyone have any problems with the article, then please just go ahead and change or even edit out the section that you think gives out bad advise. No need to ask anyone for permission. Just state a brief reason for your change is more than enough.

God knows this article can sure use a lot more contributors. Especially from experienced Arch users.

Last edited by zodmaner (2009-09-04 16:25:07)

Offline

#81 2009-09-04 16:13:06

shining
Pacman Developer
Registered: 2006-05-10
Posts: 2,043

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

sorry about that, broch's post really revolted me.

I don't know how I can ignore someone saying that kernel26-lts is useless, pacman -Sc is useless, asking to patch pacman to detect a shell command (pacman -Syy && pacman -S package), and pretending that being careful about AUR packages is worthless.

But I will now try to ignore that.
Sorry for the noise, feel free to delete all my posts related to what broch said so that this thread can go on in peace smile


pacman roulette : pacman -S $(pacman -Slq | LANG=C sort -R | head -n $((RANDOM % 10)))

Offline

#82 2009-09-04 21:21:45

Gen2ly
Member
From: Sevierville, TN
Registered: 2009-03-06
Posts: 1,529
Website

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

@shining

Hell, I didn't mind it at all.  If I did something without thinking, I'd appreciate someone telling me rather than ignoring me.  That's the thing about documentation, one person that is an expert on the subject can put together a better article than a 1000 peoples piecemeal additions.  Editors and discussion like this help keep documentation sane.  Thanks to lseubert who did a good job collecting the pieces and putting it in a coherent sense.


Setting Up a Scripting Environment | Proud donor to wikipedia - link

Offline

#83 2009-09-05 01:00:44

Allan
Pacman
From: Brisbane, AU
Registered: 2007-06-09
Posts: 11,650
Website

Re: [renamed - again, sort of] Enhancing Arch Linux Stability Wikipage

broch wrote:

Because kernel version was taken arbitrary (there is no statistical proof that this particular version is better than any stable version before or after)

You do realise that is the kernel version being supported by some kernel devs in the long term and was chosen specifically because serveral distros had based their long term support releases on it. So, you are right, no statistical proof but a large amount of anecdotal evidence...

Offline

Board footer

Powered by FluxBB