Hello all. I have a problem in that I have a failed upgrade attempt to Fedora 37. It’ll probably take two stages to ask my question thoroughly because I’m stuck in how to ask my question. After the upgrade attempt (from GNOME asking “Upgrade available! Do you wish to upgrade now?” or something close to that to which I responded in the affirmative), I got a character mode screen with the following message:
[begin message]
Generating “/run/initramfs/rdsosreport.txt”
Entering emergency mode. Exit the shell to continue.
Type “journalctl” to view systemlogs.
You might want to save “/run/initramfs/rdsosreport.txt” to a USB stick or /boot after mounting them and attach it to a bug report.
Press Enter for maintenance
(or press Control-D to continue):
[end message]
Pressing Ctl-D doesn’t continue so entering maintenance is the only option. I looked at the system logs but I’m not particularly versed in what I’m looking for so if that’s worth attaching, let me know where to get them and I’ll try to attach them to this thread.
I have a more basic problem at the moment. I’d like to copy the rdsosreport.txt to a USB stick but I’m not sure how to mount the stick from the command line. I put in a USB memory stick in the machine and I see a “sdb” and “sdb1” appear in the /dev folder but when I “ls” either one, I don’t see the preexisting file there (which I had there to verify that I had the correct device). The “sdb”s are not there when I remove the stick so I’m pretty certain that they represent the stick I’m inserting. I assume I have to mount one of them in some way? I tried mounting them with various mount commands but I don’t think I’m doing it correctly.
So, before I actually ask for help with the failure, how do I get the rdsosreport.txt onto a stick? Is that what is needed here? Or is it the system logs? If the latter, what files do I need to attach and where are they?
I apologize if this is a known process but I’ve been using Fedora since 34 and this is way beyond any problem I’ve ever had to deal with so it’s way beyond my capabilities. Thanks for the help.
Assuming what you see in /dev as /dev/sdb1 is actually the usb stick, you would mount it as sudo mount /dev/sdb1 /mnt, then to see the content ls /mnt. You have to mount a device file system (/dev/sdb1) on a mount point (/mnt) before you can see the content.
To understand how mount works use man mount and read it.
Thanks for responding so quickly, jeff V. The “sudo” command does not exist in maintenance mode but no matter, I tried variations without it. I tried “mount dev/sdb1 /mnt”, “mount dev/sdb /mnt”, “mount /dev/sdb /mnt”, and “mount /dev/sdb1 /mnt” and all ended with the message:
Thanks, Jeff V. I appreciate the sentiment and I will devote more time to it but the thing is, I was using the command properly. What I didn’t understand (in the feature deprived mode of operation I was in) is that I needed to create a folder to which to mount the usb stick to mount it to. Otherwise, I would wipe out one of the existing necessary folder in the very sparse maintenance partition that I was operating in (a Fedora 31 one from when the it was first installed).
Anyway, I was able to get the rdsosreport.txt file off of the machine. I know this is way beyond my capability to decipher. However, it’s a fairly large text file and trying to paste it in to the post violates the size limit for the forum. Any suggestions for how to get other eyes upon it? There doesn’t seem to be a way to add attachments. Thanks.
What others have done is use google drive, upload the file there, and make the file accessible. Then add a link in your post so it can be reached.
Others use similar methods with different sites, including pastebin, though pastebin has a default time limit of 24 hours for access.
A free site for uploads of this sort is hastebin, which can be easily used by most.
From the README.md for that utility is this
# Haste Client
[](http://travis-ci.org/seejohnrun/haste-client)
haste-client is a simple client for uploading data to haste-server. All you do it pipe data in STDIN:
`cat file | haste`
And once the output makes it to the server, it will print the URL to STDOUT.
This can be really really cool in combination with `pbcopy`, like:
* mac osx: `cat file | haste | pbcopy`
* linux: `cat file | haste | xsel`
after which the contents of `file` will be accessible at a URL which has been copied to your pasteboard.
## Installation
``` bash
gem install haste
```
## Making uploading file contents easier
If you supply a valid file path as argument #1 to the client, it will be uploaded:
``` bash
# equivalent
cat file | haste
haste file
```
It is installed at a user (not system) level so the executable haste is placed in ~/bin.
I went the easy (Google docs) route for now. Haste sounds interesting particularly when my Fedora box is back up.
Anyway, here’s where my rdsosreport.txt is located. If it’s of any value, I think the problem is somewhere in GNOME’s setup. It looks like GNOME tries to get started (the graphical screen comes up with the ‘please wait’ spinny going) before it collapses into the above error message).
If that’s the case, is there an incantation I can perform that gets basic machine started with say the bash shell without starting GNOME? I expect I’ll need to do that to perform whatever fixes are necessary anyway? Thanks again, and in advance, for taking a look at the report and making suggestions.
In the file I see a FAILED fsck on the root file system.
This is the tail end of that file which shows the part I thought of interest; particularly the first 15 or so entries.
If you ignore that and do not run fsck manually, then the error will spread until the file system is totally unusable. Best thing is to immediately copy any valuable files to a different disk.
Right now, I can’t run anything. I don’t mind reading up on how to fix things (with fsck, presumably) but can a general approach be suggested? What media can/should I boot to (e.g. Fedora 37 Live, make a Fedora 37 USB stick, what)? Then mount the main partition and run fsck?
If one can suggest a basic outline, I can read up on the details. Thanks.
Do not ever attempt to run fsck on a mounted file system. It may destroy the file system.
Boot to a live media then run the file system check on the unmounted file system in question. Also do not run fsck on a btrfs file system. Check what you have before doing more damage.
That seemed to do the trick. I was able to boot Fedora 36 Live DVD media. It wasn’t apparent how to get just a command line (which is what I wanted) but I got a basic core graphical I/F up from which I ran a terminal box. I fsck on all the visible partitions of the hard drive. fsck cleaned one up pretty well sending a number of fragments off and other tidbits.
Rebooting brought Gnome back to life. Shutting down after that either resumed the Fedora 37 upgrade or installed a number of updates to 37 that have occurred since 37 released. Things seem to be okay for now.
Thanks to all for the help and especially to Jeff V for walking me through the process. I’ve not had to go through this process before which just means I never had to learn how to do this. Thanks for both: helping through the fix and guiding me to learn how to do this.