| View previous topic :: View next topic |
| Author |
Message |
asturm
Developer
Joined: 05 Apr 2007
Posts: 9320
|
 Posted: Sat Jan 19, 2019 2:52 pm Post subject: |
 |
|
| He wasn't active in Gentoo in a very long time, and the last attempt wasn't particularly helpful. |
|
| Back to top |
|
 |
berferd
Tux's lil' helper
Joined: 13 May 2004
Posts: 117
|
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6069
Location: Removed by Neddy
|
| Posted: Sat Jan 19, 2019 10:51 pm Post subject: |
|
|
just pathetic ...
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20521
|
| Posted: Sun Jan 20, 2019 1:53 am Post subject: |
|
|
Yeah, it's so pathetic that they fixed it.
| excerpts wrote: | @poettering's proposal is implemented in #11443.
An implementation of the proposal #11436 (comment).
Closes #11436.
OK, that is enough for me to consider the previous behaviour documented. So I agree that we should preserve compatibility for this.
poettering commented Jan 17, 2019
BTW, for those coming late, a good fix (including a revert) has been merged now. I think all should be good now. |
Notice that the final comment by poettering was on the 17th, two days after the "taking a break" announcement.
I don't care for systemd, but this instance doesn't help the dyed-in-the-wool critics.
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
berferd
Tux's lil' helper
Joined: 13 May 2004
Posts: 117
|
| Posted: Sun Jan 20, 2019 9:32 pm Post subject: |
|
|
| Poettering in a panic tries to undo the damage his minions have done after the long-suffering maintainer for a major distro quits. Looks pretty pathetic to me. |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6069
Location: Removed by Neddy
|
| Posted: Sun Jan 20, 2019 9:35 pm Post subject: |
|
|
| pjp wrote: | Yeah, it's so pathetic that they fixed it.
| excerpts wrote: | @poettering's proposal is implemented in #11443.
An implementation of the proposal #11436 (comment).
Closes #11436.
OK, that is enough for me to consider the previous behaviour documented. So I agree that we should preserve compatibility for this.
poettering commented Jan 17, 2019
BTW, for those coming late, a good fix (including a revert) has been merged now. I think all should be good now. |
Notice that the final comment by poettering was on the 17th, two days after the "taking a break" announcement.
I don't care for systemd, but this instance doesn't help the dyed-in-the-wool critics. |
not my point. the handling it upto it was accepted was poor and unfortunately common behaviour from systemd dev's
for instance: https://www.theregister.co.uk/2017/07/28/black_hat_pwnie_awards/
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20521
|
| Posted: Mon Jan 21, 2019 2:21 am Post subject: |
|
|
I didn't claim it was free of security bugs.
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6176
Location: Dallas area
|
| Posted: Mon Jan 21, 2019 12:48 pm Post subject: |
|
|
I don't care about what systemd does, I care so little I don't even follow it's problems.
The only thing I care about is the stupidity of trying to make openrc behave and look like systemd.
It seems according to the wisdom of a certain lead developer if a systemd dev jumped off a cliff, an openrc dev should do the same.
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 655
|
| Posted: Mon Jan 21, 2019 5:51 pm Post subject: |
|
|
After the Gentoo devs, in their infinite "wisdom", followed freedesktop down this path by making it the default...
I seem to remember talking about how the "predictable and persistent" interface naming changes didn't guarantee predictability nor persistence, which would result in exactly this scenario (remote servers being inaccessible)... and how it was asinine to switch to such a system, while renaming based on MAC was a better, more consistent method for most people. |
|
| Back to top |
|
|
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
| Posted: Tue Jan 22, 2019 1:16 am Post subject: |
|
|
| saellaven wrote: | | and how it was asinine to switch to such a system, while renaming based on MAC was a better, more consistent method for most people. |
Exactly! And already available in udev. No need to screw with the kernel. |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Tue Jan 22, 2019 11:39 am Post subject: |
|
|
| saellaven wrote: | After the Gentoo devs, in their infinite "wisdom", followed freedesktop down this path by making it the default...
I seem to remember talking about how the "predictable and persistent" interface naming changes didn't guarantee predictability nor persistence, which would result in exactly this scenario (remote servers being inaccessible)... and how it was asinine to switch to such a system, while renaming based on MAC was a better, more consistent method for most people. |
For most people it was as simple as USB bus enumeration races causing their wifi adapter to change name randomly at each boot. Happens with PCI too, because a lot of laptops have theirs wired up to hotplug.
The systemd people really have no fscking understanding of the Linux kernel, or hardware. |
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3522
|
| Posted: Tue Jan 22, 2019 1:17 pm Post subject: |
|
|
| Ant P. wrote: | | saellaven wrote: | After the Gentoo devs, in their infinite "wisdom", followed freedesktop down this path by making it the default...
I seem to remember talking about how the "predictable and persistent" interface naming changes didn't guarantee predictability nor persistence, which would result in exactly this scenario (remote servers being inaccessible)... and how it was asinine to switch to such a system, while renaming based on MAC was a better, more consistent method for most people. |
For most people it was as simple as USB bus enumeration races causing their wifi adapter to change name randomly at each boot. Happens with PCI too, because a lot of laptops have theirs wired up to hotplug.
The systemd people really have no fscking understanding of the Linux kernel, or hardware. |
I've followed occasional threads where L.P. essentially says, "I looked at it for a while and didn't understand it, so..." and then goes on to re implement it his way, ignoring decades of hard-learned problems, thereby repeating them and adding some new ones.
As for "predictable, repeatable network names," and I know I'm preaching to the choir here, the ONLY thing about network devices that is predictable and repeatable is the MAC, period. (Though to be fair, I'm in the semiconductor industry, and a few weeks back someone was telling me that the Chinese companies don't really understand this and are shipping bunches of product, all with the same MAC.)
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6069
Location: Removed by Neddy
|
| Posted: Tue Jan 22, 2019 7:49 pm Post subject: |
|
|
conceptually, the predictable, repeatable network name makes sense but for some reason it isn't....
| Quote: |
Names incorporating Firmware/BIOS provided index numbers for on-board devices (example: eno1)
Names incorporating Firmware/BIOS provided PCI Express hotplug slot index numbers (example: ens1)
Names incorporating physical/geographical location of the connector of the hardware (example: enp2s0)
Names incorporating the interfaces's MAC address (example: enx78e7d1ea46da)
Classic, unpredictable kernel-native ethX naming (example: eth0) |
when I did my new build I started using this because "meh", it's not majorly intrusive, once setup its transparent. YET it has changed on my build
| Quote: | cat /etc/conf.d/net
# Set the dns_domain_lo variable to the selected domain name
config_enp30s0="dhcp"
config_eno1="dhcp"
|
the 1st entry is the one that I set during fresh install. Then some time later it became eno1 ... so much for consistency...
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54639
Location: 56N 3W
|
| Posted: Tue Jan 22, 2019 8:08 pm Post subject: |
|
|
Naib,
The implementation of predictable network naming we have swaps one set of corner cases for another, unless you have a USB doofer, then its far worse.
With USB you get a different interface name depending on the USB port its connected to.
I like the old set of corner cases that I know, so I just turned off predictable network naming, or ripped out ?dev so that there was nothing to do the renaming anyway.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6176
Location: Dallas area
|
| Posted: Tue Jan 22, 2019 8:13 pm Post subject: |
|
|
I made a udev rule (persistent) to name my net interfaces using the mac addy along with turning off predictable network naming
Never had a problem over the years.
Of course if I have to add/replace the network cards then I'll have to change the mac addy.
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3522
|
| Posted: Tue Jan 22, 2019 8:28 pm Post subject: |
|
|
| Anon-E-moose wrote: | I made a udev rule (persistent) to name my net interfaces using the mac addy along with turning off predictable network naming
Never had a problem over the years.
Of course if I have to add/replace the network cards then I'll have to change the mac addy. |
But presuming there is a proper MAC sticker somewhere on or inside the box or on the network card, with your udev urles it will be PREDICTABLE and PERSISTENT, unlike the predictable and persistent steaming pile we were handed.
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6069
Location: Removed by Neddy
|
| Posted: Tue Jan 22, 2019 8:29 pm Post subject: |
|
|
| NeddySeagoon wrote: | Naib,
The implementation of predictable network naming we have swaps one set of corner cases for another, unless you have a USB doofer, then its far worse.
With USB you get a different interface name depending on the USB port its connected to.
I like the old set of corner cases that I know, so I just turned off predictable network naming, or ripped out ?dev so that there was nothing to do the renaming anyway. |
I never said what was implemented was good, they balls it up
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Tue Jan 22, 2019 9:45 pm Post subject: |
|
|
| The sad thing is, if they'd made network devices have device nodes like every other piece of hardware, none of this would have ever happened. Nobody complains about udev's predictable names for any other subsystem because symlinks don't have all these failure modes. |
|
| Back to top |
|
|
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
| Posted: Tue Jan 22, 2019 10:50 pm Post subject: |
|
|
| Ant P. wrote: | | The sad thing is, if they'd made network devices have device nodes like every other piece of hardware, none of this would have ever happened. Nobody complains about udev's predictable names for any other subsystem because symlinks don't have all these failure modes. |
I've had the naming order of SATA drives changed too. Swapped plugs until they were right again.
For a while I experimented with an intel ethernet card and the built-in realtek. I loaded the realtek driver at boot with the other modules but not the intel module.
Later in /etc/local.d that runs last, I loaded the intel driver so it always came up as eth1
Not practical if you have a big server full of cards. Then udev does come in handy. Likewise when I re-purposed a 32 bit Gentoo machine into a router, renaming from NetX names to "lan0" and "wan0" kept me from being confused which was which. Not need for epananda type names.
Epananda's belong on a lunch menu, not network names. |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Tue Jan 22, 2019 11:07 pm Post subject: |
|
|
| For SATA, there's probably a symlink in /dev/disk/ that'd address what you were trying to do. I added a disk to my desktop recently and didn't even bother with the /dev/sd* names; partuuids are more reliable (and udev is guaranteed not to mess them up). |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6176
Location: Dallas area
|
| Posted: Wed Jan 23, 2019 12:40 am Post subject: |
|
|
| Ant P. wrote: | | For SATA, there's probably a symlink in /dev/disk/ that'd address what you were trying to do. I added a disk to my desktop recently and didn't even bother with the /dev/sd* names; partuuids are more reliable (and udev is guaranteed not to mess them up). |
I've added labels to all my sd* drives/partitions to avoid the possibility of drives changing sd order, and didn't want to deal with the long partuuids/uuids.
(my sda drive doesn't change, but I've had my usb drives change driver letters in the past)
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
Fitzcarraldo
Advocate
Joined: 30 Aug 2008
Posts: 2056
Location: United Kingdom
|
| Posted: Wed Jan 23, 2019 1:49 am Post subject: |
|
|
| Anon-E-moose wrote: | | I've added labels to all my sd* drives/partitions to avoid the possibility of drives changing sd order, and didn't want to deal with the long partuuids/uuids. (my sda drive doesn't change, but I've had my usb drives change driver letters in the past) |
Snap. I found that a USB HDD’s UUID changes if I unplug the USB HDD and plug it back in again, whereas the Label does not change. Also, Labels are a darn sight easier to remember and recognise than UUIDs or PARTUUIDs, whether or not the drive is a USB drive.
_________________
Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC systemd-utils[udev] elogind KDE on both.
My blog |
|
| Back to top |
|
|
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
| Posted: Wed Jan 23, 2019 6:13 pm Post subject: |
|
|
Tried to label my data drive. | Code: | e2label /dev/sdb1
e2label: Bad magic number in super-block while trying to open /dev/sdb1
/dev/sdb1 contains a jfs file system |
Makes sense. Is there a utility for jfs? Probably not scanned by refind anyway.
EDIT:
Found "jfs_tune -L <label> /dev/XXX" but afraid to use it as I found a warning that the equivalent ext4 command will wipe out all data on the partition.
I'd sure hat to lose 1.2TB of data. Would like to move off jfs, but don't have anywhere else to store that data.
I'm getting off-topic. Sorry. |
|
| Back to top |
|
|
Jaglover
Watchman
Joined: 29 May 2005
Posts: 8291
Location: Saint Amant, Acadiana
|
|
| Back to top |
|
|
proteusx
Guru
Joined: 21 Jan 2008
Posts: 340
|
|
| Back to top |
|
|
|
Gentoo Chat
