OMLx 4.0 Post Alpha1 improvements

omlx-4
Tags: #<Tag:0x00007fb72d621f38>
(D27) #4

Hi Rugy,

Just installed on hard drive last 2380.

Some commentaries :

  1. Launch live burned DVD:
    Everything fine …except the usual problem with OMV that no internet conexion is available in the end of launch process…
    This problem is lasting from the very first versions of OMLX3, both on my PC and my wife’s.
    The conexion exists but is not available.
    A solution is to disconnect the ethernet cable before booting the PC …and reconnect it when on the final desktop.
    On OMlx3, using “drak tools” it was posible to fix definitively the thing, I don’t know to-day without “drakconf tools”, if it will be posible…

  2. Install on hard disk from the live session :
    a- First message in Calamares : “Your equipment is not conected to internet, you will not …” => I know, I know…

b- Sugestion : Would it be posible to add a game of cards (Kpatience for instance) in the basic version, so that we can play cards while the systen is being installed? 15 minutes (no SSD here) is a long time…

c. Calamares OK with new language and keyboard options. Of course “SWAP/NO SWAP” does not appear when “personal partionning” is chosen.

d. In the very end of install, new message of calamares: “THE INSTALL HAS FAILED !
External command has not finished. Command “/usr/bin/calamares-post-script” has not finished within 30 seconds. There is a exit from the command.”
=> I gave time, nothing changed, but in fact the install was absolutely OK!

e. There is no log of calamares in /var/log! (refering point “d” above? )

f. After using the distrib 5 minutes, when everything seemed to work perfectly => FREEZE COMPLETE OF THE SCREEN! This is a recurrent problem I usually encounter when using NOUVEAU with my Nvidia card! Nothing to do but RESET.
=> The problem I see is that with previous alfa pre-version (2237 if I remember well), I had no posibility at all to finish install Nvidia driver (driver present in repos and installed with dnf)! …There was a time… we had “drakconf” and “XFdrake” to try to fix the thing …but to day, what? …!

Conclusion :
The 2280 works quite well.
Looks like all other Plasma5 distribs (no comment).
Calamares needs a bit of tuning in the end.
The new processes (DNF - no more drakconf tools - no more rpm) did not convince me at all.
If the problem of freezes with Nouveau is not due to OMV, the ones of not beeing able to install Nvidia driver, as well as the non availability at boot of internet conexion doesn’t happen with other distrib I use (ROSA, MINT, FEDORA, SUSE).

Good continuation.

(Abucodonosor) #5

Hi @D27 ,

Thx for testing the build.

Everything fine …except the usual problem with OMV that no internet conexion is available in the end of launch process…
This problem is lasting from the very first versions of OMLX3, both on my PC and my wife’s.
The conexion exists but is not available.
A solution is to disconnect the ethernet cable before booting the PC …and reconnect it when on the final desktop.

We have some switches in network stack some things got broken atm.

But is seems you have this problem on this hardware for a long time.
Does any bug report exists about this ? I would like to fix it.

When not can you open one please ? With full machine specs and at least a lspci and dmesg
output attached.

Of course “SWAP/NO SWAP” does not appear when “personal partionning” is chosen.

Well it can’t :slight_smile: It does appers only on ‘auto’ partitioning so the user does not need to
partition manually just tu turn swap off or set swap smaller size,

External command has not finished. Command “/usr/bin/calamares-post-script” has not finished within 30 seconds. There is a exit from the command.”

30 secs failed on SSD installtion … What kind Hardware is that ?

However I will set timeout to 2minutes now.

I gave time, nothing changed, but in fact the install was absolutely OK!

Yes installation is finished at this point so is fine.

There is no log of calamares in /var/log …

That is because last operation is ‘umount’ job , which copy the log over … while you used the installtion
which gave you error on cleanup script , no umount and so no log copy.

Also probelms with some Nvidia HW are known however some may be not fixable at all.
AFAIK Nvidia drivers should work fine but I don’t own Nvidia GPUs myself , not even for testing.

Also in general drak* of any sort gone in Lx4.

(Ben Bullard) #6

@D27, @abucodonosor this error always happens when user installs without an internet connection working. At least in my experience. As far as I know the only negative consequeces are the already mentioned /var/log/calamares.log not being appened to new install and this from Release Notes:

Internet Connection

Currently the Calamares Installer requires an Internet connection, but OpenMandriva Lx 4 will install just fine even without an Internet connection . It is perfectly OK to simply install as you normally would and proceed to use your new system as normal.

The install module ‘packages’ is what requires Internet, it does two things which can easily be done after the installation. This is optional if the user wishes to do so. First boot in to your new system then open Konsole and run:

sudo dnf remove calamares console-setup drakx-finish-install drakx-installer-matchbox squashfs-tools

which simply removes 5 packages used only during the installation process.

(Abucodonosor) #7

Currently the Calamares Installer requires an Internet connection,

No it just checks for one… You can remove these packages with or without internet connection.
( btw no such packages drak* :smiley: )

@ben79

Yes but @D27 telling us some bug in OM on his HW makes the NIC not working
on boot. So I want to fix that.

(Ben Bullard) #8

Why do you think this? Every OMA package is an .rpm.

(Zeffanyx) #9

Testing build 2377 for two days by now.
Found the following:

  • Calamares failed to remove the unneeded translations during the post-install stage. The window with the error description was so huge I had to hold Alt and drag it many times upside. Not critical, as the system actually got installed correctly.
  • The Install the System shortcut remained on the desktop even after installation (minor annoyance)
  • Any Qt5-based app fails to start and fires up the KCrash dialog once you add the global menubar panel on the desktop. Seems like it is critical.
  • Akonadi 18.12 is broken on OMV 4.0 By default it even cannot start (akonadictl start), but once I did “cp /etc/akonadi/*.conf ~/.config/akonadi” it could start with success. However, it is still impossible to set up an IMAP account in KMail (such as Google mail), because akonadi_imap_resource instantly crashes after you try to proceed in getting an API token from Google. Sadly, it is unusable.
  • I could not set up my NOPASSWD:ALL for sudo. I uncommented the appropriate line in the sudoers file and added myself to the wheel group. Logged out and back in, but sudo still asked me for root password.

What’s good:

  • UEFI installation is seamless and beautiful (apart from that little post-install bug)
  • DNF Dragora works like a charm (why it is not installed by default, btw?)
#10

If I’m not wrong, at the end of /etc/sudoers there is this line:

#includedir /etc/sudoers.d

The install puts there a file called 10-installer that override the main file’s settings.
You could add in this directory your personal config file called 99-perso for example.

(Abucodonosor) #11

@Zeffanyx

thx for testing

Calamares failed to remove the unneeded translations during the post-install stage.

That is fixed in git , 120secs timeout should be fine now I think.

The Install the System shortcut remained on the desktop even after installation (minor annoyance)

That is because post-script failed and packages and umount modules nerver run.

  • Any Qt5-based app fails to start and fires up the KCrash dialog once you add the global menubar panel on the desktop. Seems like it is critical.

That iso was not build or meant for tesing the installed system and was build in the middle of the
plasma , frameworks and application updates.

The iso was build to test grub and some calamares changes , so no worrie about crashes you saw
they are because mixed versions of plasma.

Also akonadi and friends problem are because the updates too.

(Abucodonosor) #12

@jclvanier

Yes you are right.

One can make changes to the file installer installed too ofc or just remove it and have a new one

(D27) #13

Hi,

To Abucodonosor and Ben.
A. Internet conexion:
In fact this conexion happens to work at boot from time to time. More frequently it doesn’t. I don’t understand why. With OMV3 and draktools, creating a new conexion, it has been working fine.
With OMV4 (from the very first cooker -I tried months ago- to this pre-alfa) I just din’t succeed having it working permanently.
The worry is that this behaviour happens only with OMV.

B. In my case, if Calamares did not create the log, it actually removed all “locales” packages not needed.

  • I confirm (here) Calamares doesn’t need internet conexion to install the distrib. …But it is posible the final message I received and the absence of log are due to lack of a conexion as says Ben.
  • My disk for trying distribs is a common hard drive SEAGATE SATA disk - 7200 rev/mn (my SSD serves only for my main distrib).

C. Ben. I know all packages are rpm, I was refering to urpmi.
With drakconf you could see (read) the details of any package; with dnf or Discover I did not see that. Besides Discover doesn’t edit the details of installation (downloading / installing) and doesn’t even have a lateral slider to help reaching bottom packages; not very handly.

lspci
https://www.casimages.com/i/190111093241259308.png.html]
Ethernet device
https://www.casimages.com/i/190111093449493913.png.html]
Conexion fail
https://www.casimages.com/i/190111093555156837.png.html]
Nvidia Graphic card
https://www.casimages.com/i/190111093654923262.png.html]

I did not succeed attaching dmesg in html or text mode, if somebody tells me how to do that…

Thanks

(Ben Bullard) #14

Post-edit: @D27 thanks for testing, we appreciate it and it does help.

Try installing dnfdragora, this is from the Release Notes:

== OMCC/drakconf and drakxtools to be replaced ==

In a major step forward OMCC/drakconf and drakxtools will be replaced with newer, more up to date, software. The first step in this process is dnfdragora available for user testing on the Lx 4.0 Alpha1 release. To test dnfdragora user will need to install it. To install dnfraora open Konsole and run:

`sudo dnf install dnfdragora`

dnfdragora is a replacement for rpmdrake. More software will be added to upcoming releases.

Discover is one of those unfortunate things that happen in Linux. It is a PackageKit frontend and like basically every PackageKit frontend before it it is :poop:. (that statement is strictly my opinion). Also in Discover documentation they are very honest about who the ‘target’ users for it are. Discover is a part of the KDE Plasma5 desktop and is not, never has been, a replacement for rpmdrake. dnfdragora is specifically written to be a replacement for rpmdrake.

Thanks, I was really confused by that. Now I understand. In the release notes it does tell about a wrapper for dnf that lets users continue to use some urpmi/urpme commands as they get used to dnf. There is also this.

Post-edit: I don’t mean to be lecturing people or “preaching” to them. If it sounds that way that is my poor writing and not intentional.

(Ben Bullard) #15

However for this little bit I am going to get behind the podium and do a little “preaching”.

OM Lx 4.0 Alpha1. This applies especially profoundly to people used to Mandriva/OpenMandriva but is meant for all users. When we say there are a lot of changes we are not kidding or using hyperbole. There are a lot of changes, so many in fact the everyone needs to read the Release Notes and Errata before testing this Operating System.

That concludes today’s homily.

(Ben Bullard) #16

Helpful hint: In this forum you can upload screen shots directly from file on your computer with the upload icon (up arrow icon).

(rugyada) #17

Hi,
html not, but plain text file (remember it must be .txt extension) you should be allowed to attach them.
Otherwise it could be user permissions related, but that should not be your case…

Yes, and also simply by drag 'n drop into editor window :wink:

(D27) #18

Thanks …had forgotten this…

keep in mind this dmesg was written straight after a hard stop, due to the screen freeze; so some lines should be altered from a normal boot.

(with up arrow icon)
dmesg.txt (90,3 Ko)

(by drag and drop)
dmesg.txt (90,3 Ko)

…will go to bed more clever…:blush:

(Ben Bullard) #19

Hey, I did’nt know the drag and drop either. :monkey:

Thanks for the tip @rugyada. Thanks for the screen shots and files @D27.

(Abucodonosor) #20

@D27

Thx for the file … Found out what it is wrong … Tbh I guessed that already…
Is nothing OM specific you hit an kernel bug ( I hit something similar in 4.17.x on one of my laptops )

Your problem is this:

ven. janv. 11 00:50:20 2019] NETDEV WATCHDOG: enp5s0 (r8169): transmit queue 0 timed out
[ven. janv. 11 00:50:20 2019] WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:461 dev_watchdog+0x1f3/0x200
[ven. janv. 11 00:50:20 2019] Modules linked in: msr af_packet nf_tables nfnetlink kvm_amd ccp snd_hda_codec_hdmi joydev rng_core kvm snd_hda_codec_realtek snd_hda_codec_gen
eric hid_generic snd_hda_intel snd_hda_codec usbhid hid irqbypass snd_hda_core snd_seq_dummy snd_hwdep snd_seq_oss snd_seq_midi_event snd_seq fam15h_power snd_seq_device snd
_pcm_oss snd_mixer_oss snd_pcm k10temp realtek r8169 libphy snd_timer snd sp5100_tco soundcore i2c_piix4 pcc_cpufreq acpi_cpufreq vboxguest ip_tables x_tables uas usb_storag
e ohci_pci crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel aes_x86_64 crypto_simd cryptd glue_helper ehci_pci ohci_hcd sr_mod evdev nouveau mxm_wm
i video i2c_algo_bit drm_kms_helper ttm xhci_pci xhci_hcd drm wmi button ehci_hcd usbcore dm_mirror dm_region_hash dm_log dm_mod sha1_ssse3 sha1_generic ipv6 crc_ccitt autof
s4                                                                                                                                                                            
[ven. janv. 11 00:50:20 2019] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.20.0-desktop-1omv4000 #1
[ven. janv. 11 00:50:20 2019] Hardware name: MSI MS-7693/970A-G43 (MS-7693), BIOS V10.3 03/28/2013

[ven. janv. 11 00:50:20 2019] RIP: 0010:dev_watchdog+0x1f3/0x200
[ven. janv. 11 00:50:20 2019] Code: 00 48 63 4d e0 eb 93 4c 89 e7 c6 05 1a 28 ae 00 01 e8 11 9b fc ff 89 d9 4c 89 e6 48 c7 c7 40 5f 13 82 48 89 c2 e8 c7 b5 83 ff <0f> 0b eb
c0 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 57 41 56                                                                                                                      
[ven. janv. 11 00:50:20 2019] RSP: 0018:ffff9b972aa03e98 EFLAGS: 00010282
[ven. janv. 11 00:50:20 2019] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[ven. janv. 11 00:50:20 2019] RDX: 0000000000040400 RSI: 00000000000000f6 RDI: 0000000000000300
[ven. janv. 11 00:50:20 2019] RBP: ffff9b97261ca480 R08: 000000000000004c R09: 0000000000000003
[ven. janv. 11 00:50:20 2019] R10: 0000000000000000 R11: 0000000000000001 R12: ffff9b97261ca000
[ven. janv. 11 00:50:20 2019] R13: 0000000000000000 R14: ffff9b972aa03ee8 R15: 0000000000000000
[ven. janv. 11 00:50:20 2019] FS:  0000000000000000(0000) GS:ffff9b972aa00000(0000) knlGS:0000000000000000
[ven. janv. 11 00:50:20 2019] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ven. janv. 11 00:50:20 2019] CR2: 00007fda9bfaf6c4 CR3: 000000011b9f2000 CR4: 00000000000406f0
[ven. janv. 11 00:50:20 2019] Call Trace:
[ven. janv. 11 00:50:20 2019]  <IRQ>
[ven. janv. 11 00:50:20 2019]  ? pfifo_fast_dequeue+0x160/0x160
[ven. janv. 11 00:50:20 2019]  call_timer_fn+0x2b/0x130
[ven. janv. 11 00:50:20 2019]  ? pfifo_fast_dequeue+0x160/0x160
[ven. janv. 11 00:50:20 2019]  run_timer_softirq+0x3cd/0x400
[ven. janv. 11 00:50:20 2019]  ? _raw_spin_lock_irq+0x15/0x30
[ven. janv. 11 00:50:20 2019]  ? __hrtimer_run_queues+0x110/0x280
[ven. janv. 11 00:50:20 2019]  __do_softirq+0x114/0x35c
[ven. janv. 11 00:50:20 2019]  irq_exit+0xd7/0xe0
[ven. janv. 11 00:50:20 2019]  smp_apic_timer_interrupt+0x74/0x140
[ven. janv. 11 00:50:20 2019]  apic_timer_interrupt+0xf/0x20
[ven. janv. 11 00:50:20 2019]  </IRQ>
[ven. janv. 11 00:50:20 2019] RIP: 0010:cpuidle_enter_state+0xbe/0x330
[ven. janv. 11 00:50:20 2019] Code: 80 7c 24 0b 00 74 17 9c 58 0f 1f 44 00 00 f6 c4 02 0f 85 51 02 00 00 31 ff e8 3e 3c 95 ff e8 d9 66 9a ff fb 66 0f 1f 44 00 00 <85> ed 0f
88 1a 02 00 00 48 b8 ff ff ff ff f3 01 00 00 48 2b 1c 24                                                                                                                      
[ven. janv. 11 00:50:20 2019] RSP: 0018:ffffffff82203e80 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff13
[ven. janv. 11 00:50:20 2019] RAX: 0000000080000000 RBX: 00000004f24ae18b RCX: 000000000000001f
[ven. janv. 11 00:50:20 2019] RDX: 00000004f24ae18b RSI: 000000001e79c300 RDI: 0000000000000000
[ven. janv. 11 00:50:20 2019] RBP: 0000000000000002 R08: 0000000000000002 R09: 0000000000022200
[ven. janv. 11 00:50:20 2019] R10: 0000002a68ce0d66 R11: ffff9b972aa21a44 R12: ffffffff822d3f18
[ven. janv. 11 00:50:20 2019] R13: ffff9b9728d1f000 R14: 0000000000000002 R15: 00000000cb9cf018
[ven. janv. 11 00:50:20 2019]  do_idle+0x24a/0x2a0
[ven. janv. 11 00:50:20 2019]  cpu_startup_entry+0x19/0x20
[ven. janv. 11 00:50:20 2019]  start_kernel+0x501/0x521
[ven. janv. 11 00:50:20 2019]  secondary_startup_64+0xa4/0xb0
[ven. janv. 11 00:50:20 2019] ---[ end trace e0736670ab6e2b11 ]---

So your NIC times out… However is an kernel bug.

Will try to find out more about in the next some days.

(rugyada) #21

Good suggestion :+1:
I passed it on to devs, almost sure they agree.

(rugyada) #22

KBackup

Screenshot_20190112_134602
Screenshot_20190112_134602.jpg1280×800 182 KB

.
KApps 18.12.1

Screenshot_20190112_205122
Screenshot_20190112_205122.jpg1280×800 174 KB

.
Note: This is ISO build ID 2381

(Ben Bullard) #23

Screenshot_20190112_151258
Screenshot_20190112_151258.jpg1920×1080 1.15 MB

Post-edit: Note: the screen-shot shows a bug that may have been < fixed >.