Mailing List Archive

From: BipinDas - Ditro <bipindas.k@ditro.com>
Date: Fri, 13 Jul 2007 15:06:27 +0530

> Dear List,
>
> I do have a server with Debian-2.6.8-12-amd64-generic kernel image on
> AMD64 processor.
> When i try to enter the command iptables -L -n -v the following error
> happens.
>
> "iptables v1.2.11: can't initialize iptables table `filter': Module is
> wrong version
> Perhaps iptables or your kernel needs to be upgraded"
>
> Please let me know how can i solve it?

The old kernels have '64bit kernel and 32bit userland' problem.
Your iptables is 32bit binary (try readelf -a `which iptables`) ? If so,
building 64bit binary of iptables will solve the problem. Upgrading the
recent kernel should also solve it, of cause.

-- Yasuyuki Kozakai

On Fri, 13 Jul 2007, Yasuyuki KOZAKAI wrote:

> From: BipinDas - Ditro <bipindas.k@ditro.com>
> Date: Fri, 13 Jul 2007 15:06:27 +0530
>
>> Dear List,
>>
>> I do have a server with Debian-2.6.8-12-amd64-generic kernel image on
>> AMD64 processor.
>> When i try to enter the command iptables -L -n -v the following error
>> happens.
>>
>> "iptables v1.2.11: can't initialize iptables table `filter': Module is
>> wrong version
>> Perhaps iptables or your kernel needs to be upgraded"
>>
>> Please let me know how can i solve it?
>
> The old kernels have '64bit kernel and 32bit userland' problem.
> Your iptables is 32bit binary (try readelf -a `which iptables`) ? If so,
> building 64bit binary of iptables will solve the problem. Upgrading the
> recent kernel should also solve it, of cause.

what kernel versions are recent enough to have this problem solved?

David Lang

From: David Lang <david.lang@digitalinsight.com>
Date: Fri, 13 Jul 2007 12:52:37 -0700 (PDT)

> > The old kernels have '64bit kernel and 32bit userland' problem.
> > Your iptables is 32bit binary (try readelf -a `which iptables`) ? If so,
> > building 64bit binary of iptables will solve the problem. Upgrading the
> > recent kernel should also solve it, of cause.
>
> what kernel versions are recent enough to have this problem solved?

I think 2.6.22. The compat layer sent into 2.6.21, but a fix was
needed and it seems not to be in 2.6.21.x yet.

-- Yasuyuki Kozakai

On Sun, 15 Jul 2007, Yasuyuki KOZAKAI wrote:

> From: David Lang <david.lang@digitalinsight.com>
> Date: Fri, 13 Jul 2007 12:52:37 -0700 (PDT)
>
>>> The old kernels have '64bit kernel and 32bit userland' problem.
>>> Your iptables is 32bit binary (try readelf -a `which iptables`) ? If so,
>>> building 64bit binary of iptables will solve the problem. Upgrading the
>>> recent kernel should also solve it, of cause.
>>
>> what kernel versions are recent enough to have this problem solved?
>
> I think 2.6.22. The compat layer sent into 2.6.21, but a fix was
> needed and it seems not to be in 2.6.21.x yet.

I'm not trying to pick on you, this is a common failing on the kernel mailing
list.

but guys, saying "Upgrading to a recent kernel should also solve it" when the
kernel version that fixes it was released within the last week is "a recent
version" it's "the latest and greatest, mostly untested version that you won't
find in any distro becouse it's so new"

at least this time the "recent version" is actually a released version, not a
point in CVS or an -rc like one of the similar discussions that came up
recently.

Thanks for the info. this is a good reason for me to skip rolling 2.6.20 and
2.6.21 into production and move to testing 2.6.22 instead.

David Lang

From: David Lang <david.lang@digitalinsight.com>
Date: Sat, 14 Jul 2007 12:40:52 -0700 (PDT)

> On Sun, 15 Jul 2007, Yasuyuki KOZAKAI wrote:
>
> > From: David Lang <david.lang@digitalinsight.com>
> > Date: Fri, 13 Jul 2007 12:52:37 -0700 (PDT)
> >
> >>> The old kernels have '64bit kernel and 32bit userland' problem.
> >>> Your iptables is 32bit binary (try readelf -a `which iptables`) ? If so,
> >>> building 64bit binary of iptables will solve the problem. Upgrading the
> >>> recent kernel should also solve it, of cause.
> >>
> >> what kernel versions are recent enough to have this problem solved?
> >
> > I think 2.6.22. The compat layer sent into 2.6.21, but a fix was
> > needed and it seems not to be in 2.6.21.x yet.
>
> I'm not trying to pick on you, this is a common failing on the kernel mailing
> list.
>
> but guys, saying "Upgrading to a recent kernel should also solve it" when the
> kernel version that fixes it was released within the last week is "a recent
> version" it's "the latest and greatest, mostly untested version that you won't
> find in any distro becouse it's so new"
>
> at least this time the "recent version" is actually a released version, not a
> point in CVS or an -rc like one of the similar discussions that came up
> recently.
>
> Thanks for the info. this is a good reason for me to skip rolling 2.6.20 and
> 2.6.21 into production and move to testing 2.6.22 instead.

Thanks for pointing out that. Indeed saying just only "recent version"
gives no information.

-- Yasuyuki Kozakai