It is not often you pick your DE and you have two of everything, one being an X11 and one being Wayland. I use KDE, and there is KDE X11 and KDE Wayland and both seem to have problem with them. In one once you enter you Secret Password, it comes back and asks you for it again and again; in the other you get to the screen which then freezes, so you go back and start all over again. I think I am now using the X11 version. I’ve also nuked SDDM and switched over to LightDM, but what is it with the X11 version and the Wayland Version. Choose one or the other, but offering 2 of everything? makes life more complicated than it needs to be.
Don’t say that ! someone’ll remove the “wrong” one … of something. …again.
That said, I don’t use Wayland because I have 2 monitors and it only lets me place the 2nd on the right when my physical monitor is on the left and it won’t allow you to select a “primary” monitor - such fundamental shortcomings you wonder how it got released.
IT today is just a huge fragemented mess and is most likely going to keep getting worse. But you can help stop this: the next time someone says “this new programming language is really cool !” just tell them to **** off … and maybe use a mallet to get your point across.
(Not having any issues with KDE at the moment.)
Thanks for the tip. I too use two monitors. I’ve have been educated that “Wayland” is still "experimental: while X11 is the “Standard” DM, still I wish that someone would make Wayland and optional inclusion. I know that it plays havoc with me. I have since ditched SDDM and installed LightDM as I do not like GDM.
Other display related things I’ve found which may be of use:
- Use VGA and DVI but not HDMI :
I have one monitor connected to the vga out on the video card - never had any problems on that.
The other was on the card’s hdmi out. There’s also a kvm (keyb/video/mouse) switch in the middle to muddy the diagnosis waters.
I was getting screen blacking out occasionally - most days for a while in the afternoons usually - on the hdmi monitor only.
There was talk online about it being related to the hdmi audio.
I couldn’t find any of the settings they were talking about and don’t currently have the time to look into what my system is using for audio - possible complications from the fact I have Gnome and KDE stuff installed etc - so for a possible quick fix I got a DVI-to-HDMI cable (£3.76 new on ebay) and put the hdmi monitor on the card’s DVI out and have had no screen problems since.
- Screen alignment
When I first switched from HDMI to DVI the monitor’s image was slighlty off and none of the “auto adjust” options were available. So there was a gap in my extended desktop where the 2 monitor images “meet”.
I manged to fix/clear this by changing both monitors to the highest resolution available on both that was just below the resolution of the lower one so where I have
I changed both to 1280x1024 where the alignment corrected itself and then back to their original resolutions where they all remained correctly aligned and no issues with that since.
- Diplay managers
I’m not having any issues with the standard Display Manager(s). I updated an issue with my install notes for KDE here :
- Login loop
I had what sounds like the same problem as you “…Secret Password, it comes back and asks you for it again and again” a while back.
Alas I don’t seem to have made any notes on how I fixed it…but I do now recall having to get selinux to relabel everything as part of fixing stuff and it would make sense that an selinux denial during login would send you back.
To fix it I had to do the emergency target thing at boot to access the file system (you could also boot from an alt OS on a usb drive and mount the filesystem) and then
create the file
/.autorelabel and reboot. I may have done
touch /.autorelable if it was already there. But you should see the system being relabeled during the next boot.
- Nvidia problems ?
My original nvidia issues - mostly solved by switching to a radeon may in part have been hdmi related but I’m not going to look into that just yet.
- KDE multi monitor desktop problem
Sometimes when I login, 1 monitor is black (no wallpaper) and does not respond to right clicks but is otherwise usable. When this happens the other monitor will have 2 wallpaper images overlapped.
You can restart KDE plasmashell without logging out/losing your workspace state and this will clear that problem - similar to restarting Windows explorer when it gets confused.
kquitapp5 plasmashell || killall plasmashell && kstart5 plasmashell
from kwin - Can I restart the KDE Plasma Desktop without logging out? - Ask Ubuntu
For persistent offenders a taser may be more effective: short,sharp,shock,they won’t do it again.
Thanks for the reply. I am still aways from finishing configuring RL 8.5, but at least it is now installed, and I keep ticking off things on my “Punch List” before I swap it out with my current WS.
One of the things will be to temporarily remove the DVI- D cable from leopard and add it to ocelot to create a 2-monitor system. The graphics card is a new / old Nvidia GeForce GTX 1050Ti, but has no VGA, so I got a Display Port => VGA adapter to hook into my 4-position KVM switch… which is VGA . The second monitor will simply plug into the DVI-D port on the card.
Om screen alignment… I have a 24" Ben Q and a 20" ViewSonic both True 8 Bit Color and non flickering – important because I am an epileptic – so I am sort of stuck at HD 1280x1024 resolution
3.I’ve got KDE 5.whatever installed 10 VD each with a separate wallpaper and its own set of widgets
Login loops. X11 version of KDE lead to login loops, switch to KDE Wayland and you could login only to have the blood screen freeze, then it was back to X11 and at last yo could login. Very Clumsy. I Nuked SDDM and switched to LightDM and that solved the problem (of course you are going to have the hassle of installing LightDM but it was worth it).
Nvidia problems. Yep!! The nouveau driver works just fine. It is currently running on openSUSE 15.3 Leap. On RL 8.5 however I eventually want to install the Nvidia drivers , but that is literally will be the LAST thing I will do, long after I have finished the configuration, have the 2 TB cleaned off and set up for BACKUP, have BackupNinja installed and configured and I have the entire disk Backed up so I can restore the entire disk just in case this goes south. I’d rather spend 3 hrs. restoring the system than 3 months because I have a jacked system that has problems.
The Multi Monitor problem has occurred to me me too. The solution – at least in CentOS 7.x – is to going to System Settings => Display Configuration and more likely than not you will see on display over lapping the other. Pull them apart so that they are edge-to-edge next to each other and that should solve the problem. I’ll find out once I temporarily set ocelot up as a 2-monitor system.
Again, Thanks for your reply.
Not convinvced about the move to LightDM. As I say, I have sddm, x11 and KDE 5(.18 on RL 8.4 and .23 on RL8.5) and no login loops so the real problem must be or have been something else - a conflict specific to your system or something that got cleared somehow about the same time you switched to LightDM. But hey ! if it’s working and you’re happy with LightDM then why not stick with it.
On the need to relabel my system - pretty sure it was after restoring files as part of trying to recover the system which’d make sense. In some situations relabeling will happen automatically and you may not even notice or it’ll look like the system has hung for a bit - that sounds like your wayland freeze and x11 retry unless it was happening repeatedly.
In system building/configuration the approach I’m using is to doc/script everything (slowwww process) and use VMs with live and backing qcow2 images so I can make/try changes and if all’s well commit to the base image and if not, discard the changes and try again. Then when all done I’ll have an image I can use to create a working system and a set of scripts I can use/tweak to configure a new build including when the next upgrade-without-an-upgrade-path arrives.
GTK ! that sounds like the right fix for the problem - I’ll try that if/when it happens next.
all the best !
Still horsing around with ocelot, but it is finally – finally – becoming an operational machine, though not without some close calls: I updated the machine and suddenly openSUSE disappeared from the RL menu, I could not get in via the RL menu or via the “backdoor” ie from openSUSE, I was forced to try and edit the menu entry and I got in, then changed SELinux to “disabled”, then switched back first to “permissive” and relabled everything, and for good measure disabled LightDM and switched back to SDDM, then rebooted the machine and crossed my fingers and up popped RL no problems!! I then changed SELinux back to “enforcing”, crossed my fingers, and rebooted.RL popped up fine. I’m using X11 and the login loop seems to be gone. openSUSE is back on the RL main menu and I can boot into it no problems as well as RL. So I am back to SDDM but am holding LightDM in reserve. I’ll be curious to see what happens when I next update the machine to see if in the middle of the update it freezes and then starts spitting out error messages that require a reboot to finish the update process (yes it had gotten that bad), but so far even that has not happened again.
I finally wiped the 2 TB clean and got it set up in fstab, so now I ONLY have RL 8.5 on the 1 TB NVMe 4.0 drive, and openSUSE 15.3 Leap on the 2" SATA drive. I got BackupNinja installed, configured, and working so I have at least some Backupups – it still needs to be fine tuned. I have yet to backup the Entire Drive using the dd command ( I goofed – I packed up the book with my cheat sheet in it. I also packed my Knoppix Disks in yet another Box). I got my Grub-Customizer-like (using GC + RL menu) to work… so far). Still need to backup all the key files ala GC. Still need to hoop up the second monitor temporarily to see if that will work with the particular nVidia card that is installed. Saw the “New Apartment” that I’ll be holed up in for the next few months before – hopefully – I’ll be allowed to move back into my newly renovated apartment. With luck I get the punch list finished and I’ll be able to backup my entire drive, and then – and ONLY then – I’ll go for the Full Monty and install the nVidia drivers. I suspect the nVidia drivers destroyed my last install, so right now I am still using the nouvue driver. If the nVidia drivers toast the drive as i suspect it might, I can then simple do an entire reinstall from the backup. But that is still months down the road. Right now I am more-or-less a happy camper that I now have an actual working machine. Still was unable to install Scribus, and could not find a way to install Cinnamon which is my alternate DE.
Stay in touch. I suspect the entire network comes down in 2.5 weeks.
No monitor issues at all here since I switched from HDMI to DVI.
I imagine you’re zipping that dd of the disk… you probably know this already but I’ll mention it just in case: clonzilla will (allegedly) make/reinstate a disk copy faster than dd as it understands filesystems and skips parts/optimises the process (IIRC).
dd-ing a disk did occur to me as part of the backup process but dd-ing disks these days can take a long long time…
I looked at various backup options but you may notice many of the online examples on making a backup omit restoring from that backup and there are issues with some of them eg. duplicity won’t write to a destination with existing files…I wasted weeks on ReaR and the inner workings of mbrs/grub for a disk layout/mbr/grub recovery option and finally dumped ReaR… well it’s still part of the backup routines, I’ve just stopped using it and will get around to removing it at some point.
I now just use rdiff-backup with some scripts and a cron job for the main nightly LVM snapshot based backups and rsync to copy that main/incremtal backup to an external for offsite rotation. Then restoring files involves either directly copying for latest or using rdiff-backup for previous versions and full disk/system recovery is a sequence of: a minimal install to get disk layout/mbr etc in place, file deletion to remove the install files, rsync to replace the backup files, UUID edits to line up the boot data (latest info stored as part of the backup) and, ./autorelable to resync the selinux labeling and it works ! (documented procedure / set of notes to hand) and is less hassle, less complicated and more reliable than ReaR (for recovery with an iso … unless you’ve got a std 1 disk setup). Though I eventually want a more robust disk layout/ mbr restoration solution.
Also noted while sorting out backups that cloud backups are mostly a non-starter given the size of current systems and typical internet upload times… so more work to do there… more work to do everywhere…always more work to do…
Took a look at Scribus: looks interesting but like you I couldn’t see any easy way to get it on RL so decided against wasting any more time on it - I’m running a Windows VM for office anyway - I know ! I know… but there is no realistic alternative to Excel. …
…that windows VM needs to be on the network to access the printer and a samba share on the host (primary workstation) BUT I don’t want windows connected to the internet otherwise except when I decide to let it try updates (on a backed qcow2 of course)…so the easiest way to do that is with the restrictions on the router (not the host) which means the VM on the LAN which led me to macvtap (kills network performance on the host) and then I stumbled across Virtual Functions (VFs) but that doesn’t work at the moment so after 3(?) days of effort I gave up on a virtual solution and ordered another NIC just for that VM - hopefully that works with a direct PCI passthrough on the qemu//session based VM - fingers crossed.
That’s where I’m at.
Good luck with the moves !
Thanks for the update Bobar. Yeah BackupNinja does – allegedly – has rsync, rdiff, and a few other things, but it basically does incremental backups weekly (or daily, hourly, or… I chose weekly). Have never relied on it to restore anything. It saves things in several formats, and I chose bzip to compress things down to save as much space on that 2 TB drive as I can. I can usually go 1.5 - 2 months before i manually go in and delete the oldest backups. It is worth a look. It is a fairly simple program to use, and you don’t need to be a programmer or engineer to set it up. The only thing that is missing on it would be a rolling delete so that after x number of backups the oldest one is automatically deleted. That said if you actually know what you are doing you could in THEORY set it up as some type of chron job, which is in fact is more or less exactly what it does: Every, say Sunday, backup y. There is no way I found that will limit the number of backups, before it automatically drops the oldest one. Your backup needs probably far outstrip my rather meager needs ( we are talking 2 whole computers here!! ) but depending where and in what capacity BackupNinja is definitely worth a look. A friend of mine who is a Fully Professor and writes tons of grants has computer needs (Windows of course) provided for by the university, and every night her computer is automatically backed up, and the backups are stored both off site, as well – I think – in the “Cloud”. She had one catastrophic computer failure and it took about 24 hours for whoever backups her machine to restore her files. I am also sure the university pays a vast sum of money to backup her research computer, she basically does not have to worry about her computers, where as I simply fly by the seat of my pants, to keep my home workstation up and running, and I have to make sure I can restore all my important files if disaster strikes.
The whole disk backup using the dd command is usually compressed in gzip format, and you’re right: It does take a lot of time to back up the entire drive, which is the major reason I only do it once a quarter. It is usually the last thing I start before I go to bed at 0 Dark Thirty, thus by the time I prop my eyelids open in the morning it has finished. The entire disk backup has rescued me a number of times. The only partition / directory that really needs to be backuped more frequently – usually – is /home. For that I find my latest BackupNija backup, uncompress it and copy any files that have changed over the past quarter. At worst I lose a weeks worth to changes, not 3 months worth of changes. For files / directories that usually do not change a lot they are usually stored on a backup computer. My genealogy research is on both my OLD (2004) computer, and that data is also on my current existing workstation (2014) leopard, and will probably will eventually be moved over to ocelot via ssh as well.
Thanks for the Best Wishes on the impending move as I am going to need it. The only good news is once I move OUT is that 90% of the stuff will remain in their boxes for the next 2-3 months, thus once I get the word I can move back IN – hopefully the same apartment as I know where everything will go down to the inch, and is configured for the disabled – it will be almost a piece of cake – the only thing that will be unpacked is the Stereo, Bed, a few clothes, pots, pans, and the computer network… assuming I do not end up in a battle with the Phone and Electric Co. over transferring my services for a 3 month move.
The biggest stress so far is on my not quite 2 Yr. old Belgian Sheepdog Puppy who is currently in training to become my 3rd Service Dog. The place looks like a disaster zone and he is probably wondering if I’ll pack him up too; now my 15.5 year old CAT… she is rather blahzay about the whole thing, having already been through at least one move two years ago, and several major changes in the household. Her philosophy is to sleep all day, demand to be fed twice a day, go to sleep, and stay out of my way. Once I turn in at night she comes over and purrrs in my ear, and snuggles up to me. Maybe because I am WARM and she wants a WARM BODY during our chilly (COLD – tonight it is suppose to get down to between 28-29 degrees and that is bloody FRIGID for Tucson) nights.
ReaR got a stay of execution… there was a simple workaround to the most recent problems and now its working its easier than the manual process.
Also, I left something out on the list of actions to manually recover a system (which may be of interest) : I had to run recovery (from RL installer) and access/unlock sda2 (LUKS) once before I could boot successfully - no idea why but others online had same thing.
Did briefly look at BackupNinja after seeing it in posts here … but I have a working, flexible, reliable setup now that does everything I currently need so sticking with that for the moment.
Got the new NIC for the VM: works perfectly but you need VT-d and have to make a change to grub.cfg for it to work and no reasonable way to get it to work under qemu://session so had to move the VM to qemu://system …
One other thing occured to me that might be an interesting experiment. I haven’t tried it but
qemu-img convert -cp -f raw -O qcow2 /dev/sda sda.qcow2
should (my reading of the docs) copy your disk to a compressed image which can be accessed as is - i.e without unzipping - from a VM. It will likely be modified in the process unless you use it as a backing file. You should then be able to convert the qcow2 file back onto a disk if need be via the reverse process.
Keep us posted. As of Tuesday or Wednesday I start taking the entire Network down, then comes a cleaning and packing of each component, and God only knows when I’ll be up again. I’m surrounded by boxes. Moving day is Feb. 23rd.
I’ll catch you all on the flip side.