apt-pinning

[nobody]

It seems the new release moved to a new system for enabling the APT repositories. Whereas in the old system all you had to do was uncomment a line, now there is some pinning system that is completely undocumented.
I find it rather customer-unfriendly that this change is forced on users without proper documentation.

So please, can someone explain how I add the debian packages so I can install NFSd?

thank you

[Cheeseboy]

Hi,

Your post is a bit confusing.
If you have views on changes in the OS you should contact Excito.
Or at least explain what "the new release" and "the old system" refers to.
You don't even let us know if it is a b2 or a b3 system...

As to the question, what is it you are trying to install? rpc.nfsd should be in the nfs-kernel-server package.
I don't think I had to change anything in /etc/apt to install it on my b3, but it was installed a long time ago before the update to 2.4.

I have a fresh b2 2.4.1 install though, and;

Code: Select all

niklas@b2:~$ sudo apt-get -s install nfs-kernel-server
[sudo] password for niklas: 
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following NEW packages will be installed:
  nfs-kernel-server
0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.
Inst nfs-kernel-server (1:1.2.2-4 Debian Squeeze subset:1.0/upstream_stable [powerpc])
Conf nfs-kernel-server (1:1.2.2-4 Debian Squeeze subset:1.0/upstream_stable [powerpc])

Cheers,

Cheeseboy

[Cheeseboy]

I'm sorry. I forgot that I did this on b2 (2.4.1):

Code: Select all

# change_distribution elvin -u

http://wiki.excito.org/wiki/index.php/C ... ces_easier

[nobody]

Indeed the version was motted. Its a B3. But I believe this issue is identical for the B2

Before the 2.4 release, all apt repos where in a file called /apt/sources.list. You could disable or enable repos by commenting out specific lines in that sources.list file.
After the 2.4 release there is no more sources.list, and it is replaced by some undocumented system of pinning. So how do I enable/disable repos in this new system now that the sources.list is no longer there?

thanks

[johannes]

Ok, sorry for not being more cleare on this, but the new system _does_ have the regular sources.list. And this sources.list is used when updating from the command line. However, with the web UI we use our own sources file, which also can be modded using the change_distribution command. This is to avoid accidentally updating (using the web UI button) to some obscure sources you once used to install a package from.

[nobody]

I'm sorry but this statement is not correct. The sources.list is removed after updating to 2.4. Other users have mentioned this in other threads as well.

Sure, if I would recreate the sources list then the apt system reads it, but how could i possibly know what repos to put in there? \

[RandomUsername]

I'm sorry but this statement is not correct. The sources.list is removed after updating to 2.4. Other users have mentioned this in other threads as well.

Not in all cases i.e. mine. My B3 is up to date and I still have my sources.list.

[nobody]

you may do so still, but wait until apt asks you to run 'apt-get autoremove' to clean old unneeded packages, and you obey. I find it unreasonable to assume that users know that they should ignore this specific instruction because it will remove this particular file.

This has been mentioned before in http://forum.excito.net/viewtopic.php?f ... 9&start=45
and specific in comment http://forum.excito.net/viewtopic.php?f ... =45#p18880 and follow-up comment http://forum.excito.net/viewtopic.php?f ... =60#p18885

[johannes]

The sources.list is not removed by our update. If you choose to do the autoremove it's up to you. THe normal use case is updating using the web interface which does not remove the sources.list. Then, installing single packets using the good old sources.list is still possible.

[nobody]

yes i know i entered the command. However the apt systems makes a strong suggestion that this command may be safely run. I really find it unreasonable to assume the user understands about this particular consequence or knows that running this command removes important files. You even had a chance to rectify that this behaviour is potentially dangerous in the thread I mentioned before, but you chose not to.

Look, the suggestion to run the apt upgrade from the command line has been mad many many times in this forum. Now that there are undocumented side-effect you stand on your hind legs and scream 'it's up to you' even after being pointed to this issue before? I find that a little harsh.

[johannes]

Hmm, I apologize for sounding harsh, that was not intended. I did the usual mistake by writing the message 30 seconds before a meeting and didn't read it through properly.

Anyways I didn't realize that you were recommended to do the autoremove, if I remember correctly (which I may not) I thought is says "you may want to run apt-get autoremove" or something in that style? Hence I never imagined anyone actually doing that. If this is the case I am sure we can re-instate the dependency on the original sources file.

Anyone more on nobody's side? GIve me a +1 and I'll add a ticket on this.

[nobody]

I thought is says "you may want to run apt-get autoremove" or something in that style? Hence I never imagined anyone actually doing that.

Really? You really find it unimaginable that a user who gets a suggestion from the trusted system, follows this suggestion? Why would anyone *not* follow it? Assuming that you believe that your users trust the stability of your product, and that they trust that the system gives accurate error messages and instructions, what reason could I possibly have to assume suggestion from Apt is dishonest? I'm not trolling, I'd really like to know how you view the line of reasoning on this.

[HnLn]

+1 for me, I usally run autoremove and I only do it cause the system suggests it

[johannes]

Really? You really find it unimaginable that a user who gets a suggestion from the trusted system, follows this suggestion? Why would anyone *not* follow it? Assuming that you believe that your users trust the stability of your product, and that they trust that the system gives accurate error messages and instructions, what reason could I possibly have to assume suggestion from Apt is dishonest? I'm not trolling, I'd really like to know how you view the line of reasoning on this.

Well, I never did it for one simple reason: Don't fix stuff that is not broken. And it's not like space is an issue on the root partition. And no, not unimaginable but not likely, but I can be wrong of course. And for those who feel confident enough to do so I don't find it unimaginable that they can create their own sources.list when needed.

In any case, ticket added, fix scheduled. Thanks for bringing it to my attention.

[Ubi]

Well, I never did it for one simple reason: Don't fix stuff that is not broken.

Thats's true, but the suggestion by Apt makes user believe that it is indeed broken. In fact similar messages are also found in the realm of RPM, but *only* when an update failed to complete or there was a corruption in the RPM database. Anyway, I think all points are made. Thanks for getting the sources.list back. Maybe superfluous, but if someone posts the content of the sources.list here that may solve the detected issue in the period until the official fix.