If you are managing 1 or 10 that are not the same, then yes, but if
one is managing 10,000+ physical linu servers with a very few
models/layouts (spanning the 5-9 year life you are dealing with) and
you up front determine the mappings by model then by pcid mapping is
great.
Prior to the pciids
On 13/04/2020 07:47, Roger Heflin wrote:
The only negative part I have seen is that on one of my desktop
systems at home removing/adding a pci card changes the PCI-busid of
the built-in network card (guessing badly designed/coded bios).
And that's exactly why the whole naming this is a PITA
e the swapped-in NIC keep the name of the swapped-out NIC.
>
> My understanding of how the device names were generated was that if you
> didn't replace a NIC with an identical model, in the exact same spot,
> you could well end up with a different device name.
>
If the PCI bus setup does no
On Tue, Apr 14, 2020 at 12:44 AM Tom Horsley wrote:
> On Tue, 14 Apr 2020 00:26:28 +0200 Tom H wrote:
>
>> The improvement's when you have a multiple NICs
>
> Right, but they should have special cased a single NIC system
> and just left the name eth0, would have avoided vast amounts
> of trouble
On Tue, 2020-04-14 at 00:26 +0200, Tom H wrote:
> The improvement's when you have a multiple NICs and swap one out. You
> no longer have to edit "/etc/udev/rules.d/.rules" in order
> to have the swapped-in NIC keep the name of the swapped-out NIC.
My understanding of how th
On 04/13/2020 04:43 PM, Tom Horsley wrote:
Right, but they should have special cased a single NIC system
and just left the name eth0, would have avoided vast amounts
of trouble for people with common desktops that only have
a single NIC.
+1 Informative
On Tue, 14 Apr 2020 00:26:28 +0200
Tom H wrote:
> The improvement's when you have a multiple NICs
Right, but they should have special cased a single NIC system
and just left the name eth0, would have avoided vast amounts
of trouble for people with common desktops that only have
a single NIC.
On Sun, Apr 12, 2020 Tom Horsley wrote:
> On Sun, 12 Apr 2020 Jack Craig wrote:
>> further reading of some RH portal docs, i decided my notion was an
>> unacceptable step backward in techs stream of forward progress.
>
> I consider the "improvement" a step backwards. Certainly for
> a desktop
On Sun, Apr 12, 2020 Roberto Ragusa wrote:
> On 2020-04-10 20:53, Jack Craig wrote:
>> on F30, i would like to swap enp4s0 to eth0.
>>
>> outside the network ifcfg* files, any other place(file) to update?
>
> Having a proper ifcfg*, including the HWADDR MAC address used to be
> enough (but be
it is not a step back for enterprise servers with 2 or more vendors
ethernet cards.
We were doing a udev rule (by pci-busid) so the right cards were in
the right places, and had (pci-busid) rules for each model so that the
names were always the same. The "new" standard makes what we (and
On Sun, 12 Apr 2020 13:33:50 -0700
Jack Craig wrote:
> further reading of some RH portal docs, i decided my notion was an
> unacceptable step backward in techs stream of forward progress.
I consider the "improvement" a step backwards. Certainly for
a desktop system with one and only one ethernet
a big thanks to those that commented on this thread.
further reading of some RH portal docs, i decided my notion was an
unacceptable step backward in techs stream of forward progress.
On 2020-04-10 20:53, Jack Craig wrote:
hi all,
on F30, i would like to swap enp4s0 to eth0.
outside the network ifcfg* files, any other place(file) to update?
Having a proper ifcfg*, including the HWADDR MAC address used to be enough
(but be sure you rebuild your initrd since renaming
On Fri, Apr 10, 2020 at 9:03 PM Tom Horsley wrote:
> On Fri, 10 Apr 2020 11:53:11 -0700 Jack Craig wrote:
>
>> on F30, i would like to swap enp4s0 to eth0.
>
> You have to get the system to stop inventing "immutable" names
> first, which means adding net.ifnames=0 biosdevname=0 to the
> kernel
On Fri, Apr 10, 2020 at 8:54 PM Jack Craig wrote:
>
> on F30, i would like to swap enp4s0 to eth0.
>
> outside the network ifcfg* files, any other place(file) to update?
>
> i scanned /etc/ find no additional references.
The simplest is to add "net.ifnames=0" to "GRUB_CMDLINE_LINUX" in
thx frank,
tom got me pointed in the right direction and your info here is a Big push
down that path.
thx again, jackc...
On Fri, Apr 10, 2020 at 12:13 PM Frank Elsner
wrote:
> On Fri, 10 Apr 2020 11:53:11 -0700 Jack Craig wrote:
> > hi all,
> >
> > on F30, i would like to swap enp4s0 to
On Fri, 10 Apr 2020 11:53:11 -0700 Jack Craig wrote:
> hi all,
>
> on F30, i would like to swap enp4s0 to eth0.
For this my /etc/default/grub contains
GRUB_CMDLINE_LINUX="net.ifnames=0 ..."
Greetings, Frank
___
users mailing list --
thanks tom, i knew there was more to it. thx, jackc...
On Fri, Apr 10, 2020 at 12:03 PM Tom Horsley wrote:
> On Fri, 10 Apr 2020 11:53:11 -0700
> Jack Craig wrote:
>
> > on F30, i would like to swap enp4s0 to eth0.
>
> You have to get the system to stop inventing "immutable" names
> first,
On Fri, 10 Apr 2020 11:53:11 -0700
Jack Craig wrote:
> on F30, i would like to swap enp4s0 to eth0.
You have to get the system to stop inventing "immutable" names
first, which means adding net.ifnames=0 biosdevname=0 to the
kernel boot parameters which means learning how to use
grub2-editenv
hi all,
on F30, i would like to swap enp4s0 to eth0.
outside the network ifcfg* files, any other place(file) to update?
i scanned /etc/ find no additional references.
also any wisdom that sez this idea is folly??
tia, jackc...
___
users mailing
20 matches
Mail list logo
