First result:
That said, I have no idea if that will help solve your problem. But if you’re running lvm for your storage, your life will be much easier if you know how to administer it.
OK guess I should have searched for some keywords then not just the documentation itself, sorry 
So I guess given this:
[liveuser@localhost-live ~]$ sudo lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
home fedora_localhost-live -wi------- 406.66g
root fedora_localhost-live -wi-a----- 50.00g
swap fedora_localhost-live -wi------- 7.90g
I would lvchange -an fedora_localhost-live/root and then try again?
I certainly believe that, but it wasn’t exactly my choice (not knowingly anyway). TBH I’d much rather have a plain old partition I can see I mean I don’t even know why I’d want an LVM to begin with so I’m a little lost here…
also funny: the bug reporter actually adds it to the existing report, although I don’t know if that’s 100% the same issue.
edit:
So I ran the command, tried again, but still got the error. Checked lvs again, and the volume is active again… 
1 Like
Yeah, idk man. Fedora installer can be finicky if you do anything out of the ordinary with it. That’s why some people prefer the more difficult, but ultimately more flexible Gentoo, Arch or BSD ways of installing. I think they just didn’t account for your use-case here unfortunately. Hope someone proves me wrong though.
It seems weird noone would have tried to overwrite a previous install tho 
Also stupid question but… can I write a new image onto a USB stick I’m currently booted from? Theoretically everything needed is in RAM, right? That way I could maybe just use F35 (the only reason I used 36 was because I thought it’s gonna release soon anyway). Cause I do not have a second PC around and I don’t know if my Windows install still works (hasn’t been booted in like 2 years).
I’ll see what the Redhat bug says in the morning, since it’s only been opened yesterday (or today depending where you live lol).
I’m wondering what would happen if I deleted the volume entirely through the partition manager (i.e. not in the installer, but before that), and created a new volume during the install…
Just don’t delete any volumes you care about
So I’m formatting an external SSD to use in a raspberry pi. The SSD is in a SATA3 to USB3 housing. The optimal I/O size is 32MB, which looks very high to me. Is this a new thing? The block size still seems to be 512 bytes. The partition boundaries are aligned to 32MB, so that’s probably OK. All my other drives are either 4K or 512?? Does anyone know if the 32MB I/O is a new thing?
Disk /dev/sdc: 223.57 GiB, 240057409536 bytes, 468862128 sectors
Disk model: Transcend
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 33553920 bytes
Disklabel type: dos
Disk identifier: 0x05eb5c4a
Device Boot Start End Sectors Size Id Type
/dev/sdc1 65535 1114094 1048560 512M c W95 FAT32 (LBA)
/dev/sdc2 1114095 468862127 467748033 223G 83 Linux
Also to answer this: the Media Writer lets me choose the USB stick as the target, but it won’t let me write to it because “Device or resource busy”.
Kinda expected that, but it was worth a shot 
The mentioned bug report got an update by the way, seems others can confirm it with reproduction steps.
So I tried this after sleeping way too damn long and it appears to be working?
I’m getting to “Installing software” now at least… so we’ll see what happens.
1 Like
An update- after all that and much hair pulling, I realized that the old SSD draws too much power to use with the rpi4. I think the rpi is limited to 1.2A max on all the USB drives. I have some newer M2 and nvme drives that use less power, so I’ll use those instead.
OK so… on the plus side: it booted
on the negative: it didn’t mount my /home, guess I’ll have to fix that (sorta expected it due to it being encrypted and all).
/edit actually I think I forgot to specify that in the installer because I was busy not forgetting any of the other partitions
Also I have package updates (obviously), and someone mentioned a while ago the GUI tools are pretty good these days… well Discover was stuck at “fetching packages” for a good 15 minutes while dnf list --upgrades took me a whole 10 seconds including fetching repos… so… there’s that.
1 Like
OK so I was curious and maybe a little stupid… went back into the installer and did the whole thing over again just to see whether it would take the home volume and actually mount it in the new install.
Well turns out it does.
Which makes me curious… how does the decryption for mounting it work? Because I didn’t have to enter any password or anything Well except for the user… is it just using that? But I thought the mounting happens before the login
Also went through my (old) dnf history for installs and from the 950-ish transactions, only 25 were actually packages I installed to really use, the rest was development stuff or random install-and-try-once installs… crazy, but nice to clean up 
Also half the transactions were from VirtualBox(?) reinstalling kernel modules every update… so yea, fun. I tried to get rid of that a while back but I couldn’t.
Just fun to see when installing 25 packages results in 589 packages being installed
OK well, so far I only found one issue:
I am getting no audio output at all, even from the audio test thingy
Why is it always audio that doesn’t work 
So I tried looking into this but I’m stumped cause I don’t really know where to start…
So to sum up:
I’m now on Fedora 36 KDE on X11 (can’t run Wayland with what I need).
The Soundchip is an ALC1220, so not exactly exotic (and worked before, so there’s that).
The audio devices are in KDE’s volume control panel, and the peak meter there also shows activity, so the sound server seems to be running, but not passing it on to the output (or it’s muted somewhere, I don’t know).
My first thought went to alsamixer because I had issues with muted output there in the past. Tried around with every possible volume there, no luck.
Checked journalctl -b and noticed:
Apr 11 11:40:45 fedora systemd[1295]: Closed pipewire-pulse.socket - PipeWire PulseAudio.
Apr 11 11:40:45 fedora systemd[1295]: Closed pipewire.socket - PipeWire Multimedia System Socket.
Not sure how important that is.
Also had something about missing pactl in there:
Apr 11 11:40:40 fedora pipewire-pulse[2134]: pw.conf: execvp error 'pactl': No such file or directory
Installed pulseaudio-utils and now that error is gone, but still no dice.
Checked Arch Wiki, and found this:
https://wiki.archlinux.org/title/PipeWire#No_Sound_or_pactl_info_shows_Failure:_Connection_refused
[tarulia@fedora]~% tree .config/pipewire/
.config/pipewire/
└── media-session.d
1 directory, 0 files
Tried restarting the unit regardless, however:
[tarulia@fedora]~% systemctl restart pipewire-pulse.service
Failed to restart pipewire-pulse.service: Unit pipewire-pulse.service not found.
That seems weird?
Also just for reference a pactl info and alsa-info:
[tarulia@fedora]~% pactl info
Server String: /run/user/1000/pulse/native
Library Protocol Version: 35
Server Protocol Version: 35
Is Local: yes
Client Index: 229
Tile Size: 65472
User Name: tarulia
Host Name: fedora
Server Name: PulseAudio (on PipeWire 0.3.49)
Server Version: 15.0.0
Default Sample Specification: float32le 2ch 48000Hz
Default Channel Map: front-left,front-right
Default Sink: alsa_output.pci-0000_0b_00.1.hdmi-stereo
Default Source: alsa_input.pci-0000_0d_00.3.analog-stereo
Cookie: 3c7c:c2d5
http://alsa-project.org/db/?f=76461a05cee5f6f4d0c741cb783ada9e7dfa9de2
Any ideas where to start? If it’s an issue with the service unit, how would I get it? I would expect it to be part of pipewire-pulseaudio (which is already installed), and it is:
[tarulia@fedora]~% dnf repoquery --list pipewire-pulseaudio
Last metadata expiration check: 0:00:46 ago on Mon 11 Apr 2022 15:42:01 CEST.
/usr/bin/pipewire-pulse
/usr/lib/.build-id
/usr/lib/.build-id/0d/a5f936580c912c25fddbb4e68170e8205af0d7.1
/usr/lib/systemd/user/pipewire-pulse.service
/usr/lib/systemd/user/pipewire-pulse.socket
/usr/share/man/man1/pipewire-pulse.1.gz
/usr/share/pipewire/pipewire-pulse.conf
But why doesn’t systemctl find it?
edit: right… systemctl restart pipewire-pulse.service --user does something, but no dice either.
Use the best tool for the job. SMB works great for a simple file share but is pretty fat. But if you need high speeds and NFS does not do this out of the box for you, then yeah, I don’t think anyone will protest this sentiment.
With that said, if you are using NTP on your network, that should suffice enough; that is what I do.
I had this issue at work the other day cold starting a machine. Solution was to slick the drives through another system and then put the drives in and run through the Fedora/RHEL installer. Who is incharge of this installer because Debian handles this better.
1 Like
It’s an issue with the F36 installer, they already marked it as a blocker for the release.
As for my audio issue… talked a bit on the Fedora Discord but didn’t get very far.
They suspected that maybe the selection in KDE isn’t working and therefore audio is still going to the default sink (the HDMI, which also doesn’t output anything). They suggested using wpctl status to get the device ID and wpctl set-default ID to change the default, but that didn’t change it either. It’s still on the HDMI and still not getting audio.
So something is very wrong here, any ideas?
I’ve already deleted/moved ~/.config/pulse, ~/.config/PulseEffects, ~/.config/pipewire, and ~/.config/pipewire-media-session with no success.
For reference, wpctl status:
[tarulia@fedora]~% wpctl status
PipeWire 'pipewire-0' [0.3.49, tarulia@fedora, cookie:1677237968]
└─ Clients:
31. uresourced [0.3.49, tarulia@fedora, pid:1628]
32. WirePlumber [0.3.49, tarulia@fedora, pid:1640]
33. WirePlumber [export] [0.3.49, tarulia@fedora, pid:1640]
54. QPulse [0.3.49, tarulia@fedora, pid:1908]
55. pipewire-pulse [0.3.49, tarulia@fedora, pid:2100]
56. Plasma PA [0.3.49, tarulia@fedora, pid:1730]
58. wpctl [0.3.49, tarulia@fedora, pid:3869]
59. Plasma PA [0.3.49, tarulia@fedora, pid:2644]
68. Chromium input [0.3.49, tarulia@fedora, pid:3689]
69. WEBRTC VoiceEngine [0.3.49, tarulia@fedora, pid:2850]
70. Chromium input [0.3.49, tarulia@fedora, pid:2883]
71. libcanberra [0.3.49, tarulia@fedora, pid:2644]
74. Chromium [0.3.49, tarulia@fedora, pid:3689]
77. xdg-desktop-portal [0.3.49, tarulia@fedora, pid:3043]
Audio
├─ Devices:
│ 39. Vega 10 HDMI Audio [Radeon Vega 56/64] [alsa]
│ 40. Family 17h (Models 00h-0fh) HD Audio Controller [alsa]
│
├─ Sinks:
│ * 41. Vega 10 HDMI Audio [Radeon Vega 56/64] Digital Stereo (HDMI) [vol: 1.00]
│ 42. Family 17h (Models 00h-0fh) HD Audio Controller Analog Stereo [vol: 1.00]
│
├─ Sink endpoints:
│
├─ Sources:
│ * 43. Family 17h (Models 00h-0fh) HD Audio Controller Analog Stereo [vol: 0.74]
│
├─ Source endpoints:
│
└─ Streams:
63. Chromium
57. output_FR > ALC1220 Analog:playback_FR
64. output_FL > ALC1220 Analog:playback_FL
96. Plasma PA
98. input_FL
99. monitor_FL
100. input_FR
101. monitor_FR
Video
├─ Devices:
│
├─ Sinks:
│
├─ Sink endpoints:
│
├─ Sources:
│
├─ Source endpoints:
│
└─ Streams:
Settings
└─ Default Configured Node Names:
0. Audio/Sink alsa_output.pci-0000_0d_00.3.analog-stereo
Edit hours later:
Progress…-ish? There is actually sound… when I turn up my speakers all the way (which would normally blow my ears out) I can just barely hear it over me typing (and it’s not a Model M).
sigh
It is the exact same issue we were discussing at length here:
That was the first thing I checked, no idea how I missed that slider… anyway, the conf file we fixed it in last time doesn’t exist (because PipeWire and not PulseAudio).
Any idea where the replacement file for that is
edit:
Found it already, it’s in the same thread posted above.
Hello everyone,
I recently got my hands on a new HP Envy x360 convertible which for some reason rfkill soft-blocks all radios whenever I flip the screen to “tablet mode”.
To try to prevent this (and because I never use it anyway), I am currently looking for a way of completely disabling rfkill, or to at least prevent it from soft-blocking.
I tried the following things:
- I disabled
rfkill,wmi,wmi_bmofandhp_wmivia/etc/modprobe.d/ - I added
rfkill.master_switch_mode=0to the kernel command line - I disabled and masked
systemd-rfkill.serviceand addedsystemd.restore_state=0to the kernel command line
I am running my own distro which is practically identical to Arch with a few quirks.
Does anyone have any idea how I can stop this from happening? Do I need to remove rfkill from my kernel or something?
1 Like
No tech support, just moral support: I have a Dell laptop that does a similar thing - if I plug in an ethernet, it disables WiFi with a “hardware switch” (there are no wifi switches on the actual laptop) I’ve never solved it, but I suspect it’s a BIOS/Hardware thing, not a software thing.
Maybe check the BIOS for a setting that jumps out?
OK so… I’m having an issue deleting a directory from my Trash for a while now and even after the reinstall it won’t go away. On the contrary, I now have a second directory I cannot delete. What’s curious is that both directories were at one point part of WINE prefixes, but they haven’t been used in ages.
When trying to “Empty Trash” in Dolphin, I only get Could not delete file /home/tarulia/.local/share/Trash/files/wine/Programs/GOG.com.
That one I deleted (i.e. moved to Trash) probably at least a year ago. The other one is the entire .PlayOnLinux directory because at one point POL wasn’t even available on Fedora cause it hadn’t been updated in forever.
Anyway, I also tried deleting through the Terminal, no luck there either:
[tarulia@fedora trash]$ pwd && ls -lah
/run/user/1000/kio-fuse-AkETtz/trash
total 8.5K
drwx------. 1 tarulia tarulia 1 Apr 16 13:12 .
drwxr-xr-x. 1 tarulia tarulia 0 Apr 16 13:01 ..
dr-xr-xr-x. 1 tarulia tarulia 4.0K Apr 16 13:01 0-.PlayOnLinux
dr-xr-xr-x. 1 tarulia tarulia 4.0K Apr 6 2019 0-wine
[tarulia@fedora trash]$ rm -rf ./*
rm: cannot remove './0-.PlayOnLinux/backup': Operation not permitted
rm: cannot remove './0-.PlayOnLinux/configurations/configurators': Operation not permitted
rm: cannot remove './0-.PlayOnLinux/configurations/function_overrides': Operation not permitted
[...]
rm: cannot remove './0-wine/Programs/GOG.com': Operation not permitted
[tarulia@fedora trash]$ sudo rm -rf ./*
[sudo] password for tarulia:
rm: cannot remove './0-.PlayOnLinux': Permission denied
rm: cannot remove './0-wine': Permission denied
Any idea of
- what’s causing this and
- how to fix this?
Probably a sticky bit set somewhere. try chmod 0777 -R trash or something like that, and see if that helps with anything.
2 Likes
… I’m confused… I haven’t touched it in a couple hours and now my Trash is empty
Oh well, if it pops up again I’ll try that, thank you.
I am preparing a cold storage back of my personal data onto an external portable USB hard drive to be stored at a friend’s house.
Obviously first task is to encrypt the drive as I do not control the security of my friend’s house and I have to assume I could lose it.
I am using cryptsetup and LUKS for this task. Cryptsetup FAQ recommend that you wipe the container prior to use. In my case it is a 4TB drive, so it is going to take a while. This where I ran across something interesting.
I first connected the drive to a USB 3.1 port and tested the write speed both for an unencrypted and encrypted partition.
Unencrypted
xx@pc:~$ dd if=/dev/zero of=/media/xx/test/foo3 bs=8k count=2000k conv=fsync
2048000+0 records in
2048000+0 records out
16777216000 bytes (17 GB, 16 GiB) copied, 139.265 s, 120 MB/s
Encrypted
xx@pc:~$ dd if=/dev/zero of=/media/xx/test/foo3 bs=8k count=2000k conv=fsync
2048000+0 records in
2048000+0 records out
16777216000 bytes (17 GB, 16 GiB) copied, 125.909 s, 133 MB/s
Seems reaosnable. Then I used dd to wipe the container and got a very slow write speed (30MB/s on average).
xx@pc:~$ sudo dd if=/dev/zero of=/dev/mapper/data status=progress
1856578048 bytes (1.9 GB, 1.7 GiB) copied, 62 s, 29.9 MB/s
This was very slow and was going to take over a day. So thought maybe it is a bottleneck with /dev/zero so I ran the following command and got (50MB/s on average)
xx@pc:~$ pv -tpreb cat /dev/zero
pv: cat: No such file or directory
6GiB 0:00:30 [54.6MiB/s] [ <=> ]
Then I ran the following command to wipe the container that doesn’t use dd expecting to get circa 50MB/s
root@pc:/home/xx# pv -tpreb cat /dev/zero > /dev/mapper/data
But it is huming along at 120-140 MB/s and should complete in about 7-8 hours.
This is what I want but I am confused by the weird bottle necks earlier using dd and just piping /dev/zero to std output.
Edit: My conlcusion is dd is very slow when writing to a block device instead of a file? Otherwise I can’t explain why dd wrote to a file at 120MB/s while slow to write to /dev/mapper/data at 30MB/s using the same source /dev/zero to the same drive
And why was /dev/zero output to std output was slow when it can clearly keep up writing to a file?
1 Like
In your first two commands you used a block size of 8k. Without that dd(I think) defaults to a block size of 512, so that might be part of the reason why the other commands were slower.
If you’re just trying to wipe the device, then I’d probably skip using count also.
2 Likes