LinuxQuestions.org
Welcome to the most active Linux Forum on the web.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 11-21-2015, 09:02 AM   #61
lems
Member
 
Registered: May 2004
Distribution: BSD
Posts: 269

Rep: Reputation: 119Reputation: 119

Quote:
Originally Posted by Philip Lacroix View Post
Hi lems. All my drives are shown both in /dev/disk/by-id and /dev/disk/by-uuid. I removed udev first, installed eudev and libgudev, then did an upgrade-all for the remaining stuff (aaa_elflibs was already up to date from the previous batch).
Thanks, seems to be something fishy with my system, then.

edit: I've reinstalled eudev and libgudev just in case, and regenerated my initrd. Now my disks show up in .../by-id as well. Phew.

Last edited by lems; 11-21-2015 at 09:07 AM.
 
1 members found this post helpful.
Old 11-21-2015, 09:23 AM   #62
GazL
LQ Guru
 
Registered: May 2008
Posts: 5,545
Blog Entries: 14

Rep: Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387Reputation: 3387
Perhaps building gpg without the usb support might be an idea in order to avoid the dependency on libudev. Anyone wanting to do fancy stuff with usb devices could always use gpg2 instead. It'd likely improve resilience and prevent similar problems in future for slackpkg users.
 
1 members found this post helpful.
Old 11-21-2015, 11:24 AM   #63
1337_powerslacker
Member
 
Registered: Nov 2009
Distribution: Slackware64-current
Posts: 802
Blog Entries: 9

Rep: Reputation: 543Reputation: 543Reputation: 543Reputation: 543Reputation: 543Reputation: 543
For anyone interested, since this thread seems to be focused on the difficulties associated with relying on an automated tool, here is how I do the upgrades:

I use an RSS program, RSSOwl, to keep me appraised of possible updates to the ChangeLog. If such shows up, I immediately go to slackware.osuosl.org to view the ChangeLog in the browser window (easier to read if the changes are long and/or extensive, such as the change from udev to eudev).

I use
Code:
 lftp -c mirror http://slackware.osuosl.org/slackware64-current/slackware64/
to download all the packages into my ~/Downloads/slackware/current/ directory. Once that's done, I cd to the directory and, using only the tools provided by pkgtools, carefully comb through the entries and use
Code:
 upgradepkg --install-new
to upgrade all listed packages in a given subdirectory. (You can specify multiple package names on the same command line). If any packages are listed as 'Removed', then I use
Code:
 removpkg <package-name>
to remove that package.

Experience has taught me not to rely totally on automated tools, such as slackpkg, to do all the upgrading for me. While I'm sure the design of slackpkg is excellent, and in over 99% of cases, works flawlessly if used as specified, there are those corner cases where the system can get totally borked. Remember the case where the mesa 32-bit package landed in the x86_64 tree by accident? I remember seeing some posts on LQ about people who did that upgrade, and the result was not pretty. I only learned about it because I downloaded the ISO image, mounted it on /mnt/dvd, and perused its contents. I came across this mistake after comparing the contents of the packages to the Changelog, and then pulled the source from the mesa website, made my own package, and installed that. Of course the mistake was fixed promptly, but a few people did get burned, and I would have been, if I hadn't seen that.

Same thing with eudev. After removing the old udev package and installing the new eudev and libgudev packages (and upgrading the eudev/udisks packages to get rid of annoying bootup messages), the upgrade went smoothly.

To some, this may seem like a lot of unnecessary work, and perhaps in 99% of cases it is, but my method works for me, and has kept my system running in tip-top shape through potentially disruptive changes in -current. I'm probably never going to trust an automated tool like slackpkg to do my upgrading for me. IMHO, it cannot replace the human eye in scrutinizing changes in the ChangeLog manually. This may ruffle some feathers, but I consider slackpkg in the same class as automated dependency resolution; that is, the potential for disaster in borking a running system is too great to rely on any automated tool. More work? Yes, but I can't argue with the results: a top-running system with very minor hiccups, and those occur rarely (monthly, if that).

Just my

Regards,

Matt
 
6 members found this post helpful.
Old 11-21-2015, 12:33 PM   #64
volkerdi
Slackware Maintainer
 
Registered: Dec 2002
Location: Minnesota
Distribution: Slackware! :-)
Posts: 1,914

Rep: Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177Reputation: 6177
Quote:
Originally Posted by kazzan View Post
Hi!
I had an issue with OpenVPN after the upgrade.

Code:
ERROR: Cannot open TUN/TAP dev /dev/net/tun: No such file or directory (errno=2)
Exiting due to fatal error
After creating the device node, OpenVPN was working again as expected.
Thanks for looking.
Tested here, and loading the tun kernel module is creating /dev/net/tun.
 
Old 11-21-2015, 04:59 PM   #65
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 3,716
Blog Entries: 3

Rep: Reputation: 481Reputation: 481Reputation: 481Reputation: 481Reputation: 481
I had absolutely no problems on all my systems.
2 months ago I took a slack64 14.1 multi-lib and started keeping up.

remember I have a ton of media stuff on them. yes those are always having to be rebuilt that has nothing to do with it.

on today 11/21/2015 I noticed a another build of gcc etc and stuff more stuff. So I
a made sure everything would up grade. drop down to level one restarted the internet.
Code:
uninstalled  sh Nvidia-#$^.run --uninstall
slackpkg update
slackpkg install-new yes I did let it upgrade my aaa_elflibs 
slackpkg upgrade-all y yes I want to wipe the multilib at that point. and other slackbuilds. 
slackpkg clean-system I want all the old stale compat32 from 3 years cleaned out. with all my builds.
run my multi-lib rsync and install
reboot 
run level 3 re-build NVidia 
and then startx.
then rebuild the 94 packages
then rebuild 21 libraries for private projects see if they work in current.
The system seems stable for what I do.
any one know of a nice patched libdv that builds in 5.2.0 let me know
other wise have to patch it my self.
KISS working fine.

Last edited by Drakeo; 11-21-2015 at 05:06 PM. Reason: aaa_elflibs
 
Old 11-21-2015, 09:27 PM   #66
TommyC7
Member
 
Registered: Mar 2012
Distribution: Slackware, CentOS, OpenBSD, FreeBSD
Posts: 503

Rep: Reputation: Disabled
I would like to also add that if anyone was doing something based on their disk uuid's, that will most likely need some editing.

For example, in lilo I had boot=/dev/disk/by-id/<disk id that points to /dev/sda> (same thing with /dev/disk/by-uuid)

There is, as far as I could find, no UUID that points to /dev/sda, /dev/sdb, etc. now.

Only UUIDs for specific partitions could be found (/dev/sda1, dev/sdb1, etc.)
 
Old 11-22-2015, 03:49 AM   #67
Didier Spaier
LQ Addict
 
Registered: Nov 2008
Location: Paris, France
Distribution: Slint64-14.2.1.2 on Lenovo Thinkpad W520
Posts: 9,475

Rep: Reputation: Disabled
Quote:
Originally Posted by bormant View Post
As I can see, there are 105 binaries in 26 packages that still linked with libudev.so.0 (from old udev package, now from aaa_elflibs), the list is attached to this message.
Fixed now, thanks Serg and Pat.
 
Old 11-22-2015, 11:19 AM   #68
navigium
Member
 
Registered: Aug 2014
Location: Switzerland
Distribution: Slackware, FreeBSD
Posts: 119

Rep: Reputation: 58
Has anyone done this upgrade on a LVM-on-LUKS System?

I tried and I still can't figure out what went wrong. The lvm commands somehow still depends on libudev.0.so, even though I did the upgrade as described. This breaks my boot process because /dev/<volumegroup>/<volume> devices don't exist. I was able to boot the system replacing the devices in fstab with /dev/mapper/<volumegroup>-<volume>. But now somehow even X fails to start, without any useful error in Xorg.0.log.

If anyone successfully upgraded a LUKS system, let me know that I know it was me who screwed something up.
 
Old 11-22-2015, 11:49 AM   #69
slacksam
Member
 
Registered: Oct 2012
Location: Germany
Distribution: Slackware, Salix, Manjaro, Arch
Posts: 161

Rep: Reputation: 24
navigium,
did you rebuild your initrd?
 
Old 11-22-2015, 12:23 PM   #70
navigium
Member
 
Registered: Aug 2014
Location: Switzerland
Distribution: Slackware, FreeBSD
Posts: 119

Rep: Reputation: 58
Quote:
Originally Posted by slacksam View Post
navigium,
did you rebuild your initrd?
*facepalm* Ok, now i feel really stupid. I just knew I forgot something.

Thank you, @slacksam!
 
Old 11-22-2015, 03:34 PM   #71
pzognar
Member
 
Registered: Jun 2015
Distribution: Debian Jessie 64
Posts: 163
Blog Entries: 9

Rep: Reputation: Disabled
Just making sure before I try anything, here is my understanding of how I must perform this upgrade on my system..,

Manually download three packages: aaa_elflibs, eudev and libgudev.
Then...

upgradpkg aaa_elflibs
installpkg eudev
installpkg libgudev

removepkg udev

Follow this with the usual update process:
slackpkg update
slackpkg install-new
slackpkg upgrade-all
slackpkg clean-system

Regenerate the initrd, because everything is in a LVM in a LUKS partition.

Run lilo.

Reboot.

Have I got that right?
 
2 members found this post helpful.
Old 11-22-2015, 04:40 PM   #72
rworkman
Slackware Contributor
 
Registered: Oct 2004
Location: Tuscaloosa, Alabama (USA)
Distribution: Slackware
Posts: 2,479

Rep: Reputation: 1083Reputation: 1083Reputation: 1083Reputation: 1083Reputation: 1083Reputation: 1083Reputation: 1083Reputation: 1083
Quote:
Originally Posted by pzognar View Post
Just making sure before I try anything, here is my understanding of how I must perform this upgrade on my system..,

Manually download three packages: aaa_elflibs, eudev and libgudev.
Then...

upgradpkg aaa_elflibs
installpkg eudev
installpkg libgudev

removepkg udev

Follow this with the usual update process:
slackpkg update
slackpkg install-new
slackpkg upgrade-all
slackpkg clean-system

Regenerate the initrd, because everything is in a LVM in a LUKS partition.

Run lilo.

Reboot.

Have I got that right?
Yep.
 
Old 11-23-2015, 06:33 AM   #73
pzognar
Member
 
Registered: Jun 2015
Distribution: Debian Jessie 64
Posts: 163
Blog Entries: 9

Rep: Reputation: Disabled
TY for the answers to my questions.

Ran the upgrade process. eudev appears fine. Can boot, can log in, can get into x-windows, can get wifi. Had the expected GPG error during the "slackpkg update" part. So, overall a success.

The only unexpected bit, as no one has mentioned it in this thread, was a warning about the md5sum for kernel-firmware-20151112git-noarch-1.txz during the "slackpkg upgrade all" part.

p.s. Virtualbox and bumblebee complained about missing kernel modules during boot. That was expected. What was not expected was that after reinstalling both, the messages still happen. This last should probably go to its own thread.

Last edited by pzognar; 11-23-2015 at 08:13 AM.
 
Old 11-23-2015, 07:14 AM   #74
bassmadrigal
LQ Guru
 
Registered: Nov 2003
Location: West Jordan, UT, USA
Distribution: Slackware
Posts: 7,319

Rep: Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027Reputation: 5027
Quote:
Originally Posted by TommyC7 View Post
I would like to also add that if anyone was doing something based on their disk uuid's, that will most likely need some editing.

For example, in lilo I had boot=/dev/disk/by-id/<disk id that points to /dev/sda> (same thing with /dev/disk/by-uuid)

There is, as far as I could find, no UUID that points to /dev/sda, /dev/sdb, etc. now.

Only UUIDs for specific partitions could be found (/dev/sda1, dev/sdb1, etc.)
I believe this is expected. This is the same behavior that exists on 14.1. UUIDs are only generated for formatted partitions[1], not disks. I have to reference my /dev/sda disk in lilo.conf with boot = /dev/disk/by-id/ata-Crucial-CT480M500SSD1_13440956798B since no UUID exists for the disk (I use UUIDs for all the other root sections in there and in my fstab).

1. These identifiers are generated by filesystem utilities (e.g. mkfs.*) when the partition gets formatted and are designed so that collisions are unlikely. (SOURCE)

Last edited by bassmadrigal; 11-23-2015 at 07:16 AM.
 
Old 11-23-2015, 10:58 AM   #75
pchristy
Member
 
Registered: Oct 2012
Location: UK
Distribution: Slackware
Posts: 504

Rep: Reputation: Disabled
Just done the upgrade to eudev on my laptop (a recent install of 64-current) before attempting it on my main desktop. Following the instructions above (thanks to pzognar and Robby Workman) all has gone smoothly bar two minor hiccups.

The first isn't directly related to the eudev upgrade, but re-appeared as a result of the "slackpkg --upgrade-all". On both my desktop and, more importantly my laptop, I've always received a complaint from kde that it couldn't find a power management back-end since upower was upgraded to 0.9.23 some time ago. The solution has been simply to recompile upower from the source package on my machine. I don't have to change anything in the slackbuild - simply recompiling it provides a complete fix! Without this fix, I get no battery monitoring on my laptop, and kde error messages on both machines when logging in. Of course, my upgrade got over-written during the update, and the error complaints re-appeared. Re-compiling has once again provided an instant fix.

The second issue only appears when I use the generic slackware kernel (4.1.13) and may be related to this thread:

http://www.linuxquestions.org/questi...gs-4175559295/

I've created an "ath9k.conf" file in /etc/modprobe.d to set the correct regulatory domain for my wifi card. If I boot my "home-rolled" kernel (4.3.0), there are no problems. If I boot the slackware generic kernel in multi-user mode (3), the regional code does not get set, and defaults to "00". During the boot, I see several calls like this:
Code:
cfg80211: Calling CRDA to update world regulatory domain
It does this several times, but stops on reaching a login.

However, if I boot the generic kernel into single user mode (1), the system seems to get stuck in a loop, continually churning out the cfg80211 message shown above, and not stopping after login!

This makes it quite difficult to use the terminal!

None of these are showstoppers, more minor irritations, as all works well on my own kernel.

--
Pete
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
How much is eudev functionally behind udev bstaletic Gentoo 3 10-30-2014 10:19 AM
LFS flushes systemd-udev down the toilet, replaces with eudev ReaperX7 Linux From Scratch 6 03-25-2014 09:38 PM
LXer: Gentoo Team Isolates Udev from Systemd (eudev) LXer Syndicated Linux News 0 05-02-2013 07:12 AM
LXer: Udev fork is a training project say eudev developers LXer Syndicated Linux News 0 02-03-2013 02:21 PM
Switching to udev from devfs kushalkoolwal Debian 3 09-02-2005 04:15 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 02:23 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration