Bug#324189: Vlan scripts are buggy

[Loic Minier]

found 324189 1.8-1
notfound 324189 1.9-1
tags 324189 + patch
thanks

Hi,

On Tue, Nov 29, 2005, Clement Hermann (nodens) wrote:
> Maybe it would be a good idea to keep this bug open until sarge support
> is dropped ? It would be confusing for sarge users otherwise (the bug is
> not resolved from their view point). But I agree the bug is fixed in
> current sid/etch version. What about the wontfix tag ? Or maybe I am
> misunterpreting the "Resolved bug section" implications ?
> In any case, it's important that the bug and its fix appears clearly
> somewhere...

 People can request bugs by distribution, eg:
  
  (this mail should help the BTS classification a bit more)

 The problem is that the bug ain't truly "resolved", but instead entered
 the "archiving process", where if no activity appears for 30 days, it
 will be locked to read-only (until reopened).

 Anyway, there are not so many reports against sarge, so that should be
 readable.

   Bye,
-- 
Loïc Minier <[EMAIL PROTECTED]>

Bug#324189: Vlan scripts are buggy

[Clement Hermann (nodens)]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Hi,

I also have this problem (a vlan is never brought down if the interface
is named vlan), on sarge machines. It looks like $IF_VLAN_RAW_DEVICE
is empty.

The problem does not occur on version 1.8-2.

I suggest this bug be tagged "sarge".

Attached is the quick (and most probably unclean) workaround I use on
sarge machines.

Regards,
- --
Clément 'nodens' Hermann
- - "L'air pur ? c'est pas en RL, ça ? c'est pas hors charte ?"
Jean in L'Histoire des Pingouins, http://tnemeth.free.fr/fmbl/linuxsf/

Vous trouverez ma clef publique sur le serveur public pgp.mit.edu.
Please find my public key on the public keyserver pgp.mit.edu.
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDjDrB0yQ2guvROZ0RAvHoAJ900cbA/eMwbwrJdIBu0xbQAduvzgCfam3E
txV0oDyUPLoRC2fRQQ01Hnw=
=t+N5
-END PGP SIGNATURE-
--- /root/tmp/etc/network/if-post-down.d/vlan   2004-08-19 11:12:22.0 
+0200
+++ if-post-down.d/vlan 2005-11-29 11:59:11.0 +0100
@@ -11,6 +11,10 @@
eth*.*)
IF_VLAN_RAW_DEVICE=`echo $IFACE|sed 
"s/\(eth[0-9][0-9]*\)\..*/\1/"`
;;
+vlan*)
+   IF_VLAN_RAW_DEVICE=`echo $IFACE|sed "s/vlan0*//"`
+   ;;
+
*)  exit 0
;;
 esac

Bug#324189: Vlan scripts are buggy

[Sylvain Coutant]

>  Could you copy your failing configuration?

It's fairly old, but I will try to dig this out and check exactly why I 
suggested to do it this way.

Bug#324189: Vlan scripts are buggy

[Loic Minier]

tags 324189 + unreproducible
thanks

Hi,

On sam, aoû 20, 2005, Sylvain COUTANT wrote:
> There is a typo line 12, one should read :
> VLANID=`echo $IFACE|sed "s/vlan*//"`
> VLANID=`echo $IFACE|sed "s/vlan0*//"`

 Actually, after careful reading, this is not a problem:
% echo vlan0012 | sed "s/vlan0*//"
12
% echo vlan12 | sed "s/vlan0*//"
12
 Your suggestion would work:
% echo vlan0012 | sed "s/vlan*//"
0012
% echo vlan12 | sed "s/vlan*//"
12
 But it is an incorrect regexp:
% echo vla012 | sed "s/vlan*//"
012
% echo vla12 | sed "s/vlan*//"
12

 I think the current is correct, but doesn't check for errors.

> /etc/network/if-post-down.d/vlan:
> Vlan with names vlan0* are not handled here so they are brought up but never 
> down.

 They are handled, there's no need for a special test on the interface
 name, that's all.  Calling "vconfig rem vlan12" is enough.  I agree
 that the IF_VLAN_RAW_DEVICE setting in eth*.* is weird, and it could be
 rearranged.

 I couldn't reproduce your bug with the following in my interfaces:

iface eth0.0002 inet manual
up ifconfig $IFACE up
down ifconfig $IFACE down

iface eth0.12 inet manual
up ifconfig $IFACE up
down ifconfig $IFACE down

iface vlan12 inet manual
vlan_raw_device eth0
up ifconfig $IFACE up
down ifconfig $IFACE down

iface vlan0012 inet manual
vlan_raw_device eth0
up ifconfig $IFACE up
down ifconfig $IFACE down

 You can try this with:
# for i in eth0.0002 eth0.12 vlan12 vlan0012; do \
ifup $i && sleep 1 && ifdown $i; \
done

 Could you copy your failing configuration?

   Cheers,

-- 
Loïc Minier <[EMAIL PROTECTED]>

Bug#324189: Vlan scripts are buggy

[Sylvain Coutant]

>  Instead, the information on which VLAN this interface should be added
>  in should appear in an additional keyword.

I agree. Just my point was not to change the whole, just to correct what seems 
to be a typo.


Regards,
--
Sylvain

ADVISEO
http://www.adviseo.fr/
http://www.open-sp.fr/

Bug#324189: Vlan scripts are buggy

[Loic Minier]

Hi,

On sam, aoû 20, 2005, Sylvain COUTANT wrote:
> There is a typo line 12, one should read :
> VLANID=`echo $IFACE|sed "s/vlan*//"`
> And not
> VLANID=`echo $IFACE|sed "s/vlan0*//"`
> As it is the case actually.

 I agree that this is a problem, but I think vlan should not use the
 interface name to guess about the use of the interface, especially when
 one can rename interfaces at will with ifrename.

 Instead, the information on which VLAN this interface should be added
 in should appear in an additional keyword.

   Cheers,
-- 
Loïc Minier <[EMAIL PROTECTED]>

Bug#324189: Vlan scripts are buggy

[Sylvain COUTANT]

Package: vlan
Version: 1.8-1


/etc/network/if-pre-up.d/vlan:

There is a typo line 12, one should read :
VLANID=`echo $IFACE|sed "s/vlan*//"`

And not
VLANID=`echo $IFACE|sed "s/vlan0*//"`
As it is the case actually.



/etc/network/if-post-down.d/vlan:

Vlan with names vlan0* are not handled here so they are brought up but never 
down.


--
Sylvain COUTANT