Re: CentOS 6/7

2018-12-04 Thread Brian Cluff

On 12/4/18 1:50 PM, Snyder, Alexander J wrote:
How does the predictable naming conventions work in VMs? I see they 
always differ slightly.
I don't have much experience with VMs and the new naming convention, but 
I would guess that it would name them based on the virtual bus that the 
system creates for the VM and would appear very similar to the native 
system.
Are we not using a pool of expected syntax like 1/2/3 ??? Or is it 
built on things like mb.vendor, # nics, and other arbitrary things 
like that?
I forgot to include what the system uses for the actual names, but I 
added to a reply to my own message... you've probably seen it by now, 
but yes it's based on PCI slot, usb port ext.  It even takes into 
account when it's a single card with multiple network devices on board.
I've been reconfiguring servers that were vMotioned from one DC to 
another ... the nic names are ALWAYS *slightly* different.
That's probably because the ethernet devices are hanging off slightly 
different slots on the motherboards.

Im not sure why the way you access single user mode had to change.
I'm not either.  Seems like they should have added something that would 
catch the word single and remap it to it's systemd equivalent.  Typing 
single, while a fairly popular way of getting into single user mode, 
wasn't a universal way to get there on every system.  While 
"systemd.unit=rescue" will work on any systemd system and the 
init=/bin/bash will work on everything and always has, but only starts 
the system with just bash running leaving you to do everything else by 
hand.  The rescue mode would be better if you want networking and stuff 
like that to just work.


Brian Cluff
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: CentOS 6/7

2018-12-04 Thread Snyder, Alexander J
How does the predictable naming conventions work in VMs? I see they always
differ slightly.

Are we not using a pool of expected syntax like 1/2/3 ??? Or is it built on
things like mb.vendor, # nics, and other arbitrary things like that?

I've been reconfiguring servers that were vMotioned from one DC to another
... the nic names are ALWAYS *slightly* different.

Im not sure why the way you access single user mode had to change.

Thanks,
Alexander.

Sent from my Samsung Galaxy S8+

On Tue, Dec 4, 2018, 13:41 Brian Cluff  On 12/4/18 12:40 PM, Snyder, Alexander J wrote:
> > Does anyone know why networking devices aren't eth0/1/2/3 but are now
> > ens0f0/enp0d0.
>
> Those are the new "Predictable Network Interface Names" based on where
> they are physically plugged into the system.
> You can read all about them here:
>
> https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/
> In a nut shell they make sure that you device names don't ever change
> which is really annoying when you have a server is hosted out of state
> and can't get your hands on it to fix all your scripts that use eth0 as
> your internet facing interface when it suddenly switches places with
> eth1 or gets bumped to eth3 for any number of reasons.
> The new names are a little jarring at first, but if you allow yourself
> to get used to them then you will be able to tell someone, without any
> doubt, based on the devices name which exact device needs to be swapped
> out when it fails, and the replacement device will be given the same
> name, as long as it's plugged into the same place)
>
> > Also getting into single user mode now is (IMHO) unnecessarily
> > complicated (typing 'single' versus now 'init=/sysroot/bin/bash').
>
> init=/sysroot/bin/bash has always worked and is my preferred way of
> getting into a system without running anything else.  That line simply
> tells grub to bypass starting the system's init system which is systemd
> on the newer systems and to instead run bash as linux's init.
>
> If you are looking to get into a true single user mode you will probably
> want to instead use one of the following lines where you were previously
> using init=/sysroot/bin/bash:
>
> systemd.unit=rescue.target (can be shortened to systemd.unit=rescue)
> or
> systemd.unit=emergency.target (can be shortened to systemd.unit=emergency)
>
> The emergency target is the most minimal of the 2.
>
> Brian Cluff
>
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: CentOS 6/7

2018-12-04 Thread Brian Cluff
Here's the description of the naming convention that is used, taken from 
the kernel source:


 Two character prefixes based on the type of interface:
   en — Ethernet
   ib — InfiniBand
   sl — serial line IP (slip)
   wl — wlan
   ww — wwan

Type of names:
b
            — BCMA bus core number
c
            — bus id of a grouped CCW or CCW device, with all leading 
zeros stripped [s390]

o[n|d]
            — on-board device index number
s[f][n|d]
            — hotplug slot index number
x
            — MAC address
[P]ps[f][n|d]
            — PCI geographical location
[P]ps[f][u][..][c][i]
            — USB port number chain
v
            — VIO slot number (IBM PowerVM)
ai
            — Platform bus ACPI instance id

All multi-function PCI devices will carry the [f] number in the
 device name, including the function 0 device.

 SR-IOV virtual devices are named based on the name of the parent 
interface,

 with a suffix of "v", where  is the virtual device number.

When using PCI geography, The PCI domain is only prepended when it is not 0.

For USB devices the full chain of port numbers of hubs is composed. If the
 name gets longer than the maximum number of 15 characters, the name is not
 exported.
 The usual USB configuration == 1 and interface == 0 values are suppressed.

 PCI Ethernet card with firmware index "1":
   ID_NET_NAME_ONBOARD=eno1
   ID_NET_NAME_ONBOARD_LABEL=Ethernet Port 1

 PCI Ethernet card in hotplug slot with firmware index number:
   /sys/devices/pci:00/:00:1c.3/:05:00.0/net/ens1
   ID_NET_NAME_MAC=enx0466
   ID_NET_NAME_PATH=enp5s0
   ID_NET_NAME_SLOT=ens1

 PCI Ethernet multi-function card with 2 ports:
   /sys/devices/pci:00/:00:1c.0/:02:00.0/net/enp2s0f0
   ID_NET_NAME_MAC=enx78e7d1ea46da
   ID_NET_NAME_PATH=enp2s0f0
   /sys/devices/pci:00/:00:1c.0/:02:00.1/net/enp2s0f1
   ID_NET_NAME_MAC=enx78e7d1ea46dc
   ID_NET_NAME_PATH=enp2s0f1

 PCI wlan card:
   /sys/devices/pci:00/:00:1c.1/:03:00.0/net/wlp3s0
   ID_NET_NAME_MAC=wlx0024d7e31130
   ID_NET_NAME_PATH=wlp3s0

 PCI IB host adapter with 2 ports:
   /sys/devices/pci:00/:00:03.0/:15:00.0/net/ibp21s0f0
   ID_NET_NAME_PATH=ibp21s0f0
   /sys/devices/pci:00/:00:03.0/:15:00.1/net/ibp21s0f1
   ID_NET_NAME_PATH=ibp21s0f1

 USB built-in 3G modem:
/sys/devices/pci:00/:00:1d.0/usb2/2-1/2-1.4/2-1.4:1.6/net/wwp0s29u1u4i6
   ID_NET_NAME_MAC=wwx028037ec0200
   ID_NET_NAME_PATH=wwp0s29u1u4i6

 USB Android phone:
/sys/devices/pci:00/:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/net/enp0s29u1u2
   ID_NET_NAME_MAC=enxd626b3450fb5
   ID_NET_NAME_PATH=enp0s29u1u2

 s390 grouped CCW interface:
  /sys/devices/css0/0.0.0007/0.0.f5f0/group_device/net/encf5f0
  ID_NET_NAME_MAC=enx026d3c0a
  ID_NET_NAME_PATH=encf5f0

Brian Cluff

On 12/4/18 1:41 PM, Brian Cluff wrote:

On 12/4/18 12:40 PM, Snyder, Alexander J wrote:
Does anyone know why networking devices aren't eth0/1/2/3 but are now 
ens0f0/enp0d0.


Those are the new "Predictable Network Interface Names" based on where 
they are physically plugged into the system.

You can read all about them here:
https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ 

In a nut shell they make sure that you device names don't ever change 
which is really annoying when you have a server is hosted out of state 
and can't get your hands on it to fix all your scripts that use eth0 
as your internet facing interface when it suddenly switches places 
with eth1 or gets bumped to eth3 for any number of reasons.
The new names are a little jarring at first, but if you allow yourself 
to get used to them then you will be able to tell someone, without any 
doubt, based on the devices name which exact device needs to be 
swapped out when it fails, and the replacement device will be given 
the same name, as long as it's plugged into the same place)


Also getting into single user mode now is (IMHO) unnecessarily 
complicated (typing 'single' versus now 'init=/sysroot/bin/bash').


init=/sysroot/bin/bash has always worked and is my preferred way of 
getting into a system without running anything else.  That line simply 
tells grub to bypass starting the system's init system which is 
systemd on the newer systems and to instead run bash as linux's init.


If you are looking to get into a true single user mode you will 
probably want to instead use one of the following lines where you were 
previously using init=/sysroot/bin/bash:


systemd.unit=rescue.target (can be shortened to systemd.unit=rescue)
or
systemd.unit=emergency.target (can be shortened to 
systemd.unit=emergency)


The emergency target is the most minimal of the 2.

Brian Cluff

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss


---

Re: CentOS 6/7

2018-12-04 Thread Brian Cluff

On 12/4/18 12:40 PM, Snyder, Alexander J wrote:
Does anyone know why networking devices aren't eth0/1/2/3 but are now 
ens0f0/enp0d0.


Those are the new "Predictable Network Interface Names" based on where 
they are physically plugged into the system.

You can read all about them here:
https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/
In a nut shell they make sure that you device names don't ever change 
which is really annoying when you have a server is hosted out of state 
and can't get your hands on it to fix all your scripts that use eth0 as 
your internet facing interface when it suddenly switches places with 
eth1 or gets bumped to eth3 for any number of reasons.
The new names are a little jarring at first, but if you allow yourself 
to get used to them then you will be able to tell someone, without any 
doubt, based on the devices name which exact device needs to be swapped 
out when it fails, and the replacement device will be given the same 
name, as long as it's plugged into the same place)


Also getting into single user mode now is (IMHO) unnecessarily 
complicated (typing 'single' versus now 'init=/sysroot/bin/bash').


init=/sysroot/bin/bash has always worked and is my preferred way of 
getting into a system without running anything else.  That line simply 
tells grub to bypass starting the system's init system which is systemd 
on the newer systems and to instead run bash as linux's init.


If you are looking to get into a true single user mode you will probably 
want to instead use one of the following lines where you were previously 
using init=/sysroot/bin/bash:


systemd.unit=rescue.target (can be shortened to systemd.unit=rescue)
or
systemd.unit=emergency.target (can be shortened to systemd.unit=emergency)

The emergency target is the most minimal of the 2.

Brian Cluff

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: CentOS 6/7

2018-12-04 Thread Ed
On Tue, Dec 4, 2018 at 12:40 PM Snyder, Alexander J
 wrote:
>
> Why did things have to change so dramatically between CentOS 6 and 7?
>
because VMS from DEC has faded from current memory

> Does anyone know why networking devices aren't eth0/1/2/3 but are now 
> ens0f0/enp0d0.
>
because interface names are now deterministic, sort of.  Before with
multiple NICs it was a race to see which became eth0, eth1 etc

> Also getting into single user mode now is (IMHO) unnecessarily complicated 
> (typing 'single' versus now 'init=/sysroot/bin/bash').
>
sorry about that

> Im not sure what to even Google to get the explanation im seeking, so I 
> thought I'd ask the brain trust!
>
Look to the Freedesktop.org section - and read the systemd for
Administrators Blog Series at 0pointer.de

> Thanks,
> Alexander.
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: CentOS 6/7

2018-12-04 Thread Herminio Hernandez Jr.
A lot of the changes are to the migration to Systemd as the init system

Sent from my iPhone

> On Dec 4, 2018, at 1:25 PM, Bob Elzer  wrote:
> 
> It's called progress.
> 
> but I agree it was a shock that first time booting. To me the biggest change 
> was going to systemd, but I actually like it better and it boots a whole lot 
> faster.
> 
> as for the network device names, I believe the names correspond to where the 
> device is on the motherboard, it really makes it hard to preconfigure it when 
> you don't know what it's going to be called that first time.
> 
> The one thing that annoys me is they changed the order of the new service 
> command it was
> service name  command ie service httpd restart
> 
> now it's. systemctl restart httpd
> 
> I don't know why they flipped the order.
> 
> 
>> On Tue, Dec 4, 2018, 12:40 PM Snyder, Alexander J > wrote:
>> Why did things have to change so dramatically between CentOS 6 and 7?
>> 
>> Does anyone know why networking devices aren't eth0/1/2/3 but are now 
>> ens0f0/enp0d0.
>> 
>> Also getting into single user mode now is (IMHO) unnecessarily complicated 
>> (typing 'single' versus now 'init=/sysroot/bin/bash').
>> 
>> Im not sure what to even Google to get the explanation im seeking, so I 
>> thought I'd ask the brain trust!
>> 
>> Thanks,
>> Alexander.
>> ---
>> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
>> To subscribe, unsubscribe, or to change your mail settings:
>> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: CentOS 6/7

2018-12-04 Thread Bob Elzer
It's called progress.

but I agree it was a shock that first time booting. To me the biggest
change was going to systemd, but I actually like it better and it boots a
whole lot faster.

as for the network device names, I believe the names correspond to where
the device is on the motherboard, it really makes it hard to preconfigure
it when you don't know what it's going to be called that first time.

The one thing that annoys me is they changed the order of the new service
command it was
service name  command ie service httpd restart

now it's. systemctl restart httpd

I don't know why they flipped the order.


On Tue, Dec 4, 2018, 12:40 PM Snyder, Alexander J  Why did things have to change so dramatically between CentOS 6 and 7?
>
> Does anyone know why networking devices aren't eth0/1/2/3 but are now
> ens0f0/enp0d0.
>
> Also getting into single user mode now is (IMHO) unnecessarily complicated
> (typing 'single' versus now 'init=/sysroot/bin/bash').
>
> Im not sure what to even Google to get the explanation im seeking, so I
> thought I'd ask the brain trust!
>
> Thanks,
> Alexander.
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

CentOS 6/7

2018-12-04 Thread Snyder, Alexander J
Why did things have to change so dramatically between CentOS 6 and 7?

Does anyone know why networking devices aren't eth0/1/2/3 but are now
ens0f0/enp0d0.

Also getting into single user mode now is (IMHO) unnecessarily complicated
(typing 'single' versus now 'init=/sysroot/bin/bash').

Im not sure what to even Google to get the explanation im seeking, so I
thought I'd ask the brain trust!

Thanks,
Alexander.
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss