Good that it worked. Luckily it didn't affect an extremely crucial package. I wonder if removing is really that needed or if you could directly reinstall the package with --overwrite; my method would be less dangerous if so. Well, I guess it's too late to try.
Yeah, sometime quoting is actually not needed.
Yeah, it happens sometime, but these days it has been really crappy. I don't recall actually having that much difficulty before, although it looks like it isn't as much as a disaster than two months ago.
Pretty good recommendation for everyone that use a computer in general.
Sadly it needs to be re-iterated nearly every major update.
You just know people are going to have issues, have no backups, claim this is their only work / school / uber critical system ... and come on here and blame Manjaro, when backups could "save their day".
Anyway ... update went smoothly for me ... KDE Plasma with full encryption.
You know what's weird? I can't find dnsmasq anywhere in the pacman.log, even though v2.80-3 was included in this update set. This must be newer behavior with pacman -- I'm assuming if something is about to go south in a hurry (like x crashing), pacman (or pamac cli, I'm not sure) immediately terminates the package being updated with a "Terminated" prefix -- at the expense of installing a "ghost" package...in my case "terminateddnsmasq."
So, finally updated 3 VMs, a laptop and a desktop PC. All graphically with Pamac GUI.
...Went trough it without any problem at all on every systems. The only "problem" is that at first shutdown/reboot, you have to wait 2 minutes because of "A stop job is running for User Manager for UID 1000". It doesn't do that after, so not a big deal at all IMO.
Well, this is pleasant. It's going to need coffee, lots of coffee...
My Testing branch KDE is a VM. Once i'd updated the mirrors & fetched all 170 updates [in the Pamac GUI], i then closed Pamac [ie, did NOT proceed with the update], logged out, changed VM to tty2, then ran pamac upgrade. After it completed [i was elsewhere, not watching it], with no apparent errors on-screen, i issued reboot. Oh dear.
Here's the end part of the journal:
Here's the mess after trying pacman -Syyu, in the tty2:
Yesterday's update of my Unstable VM went just fine, so i really did not expect today's Testing disaster. I assume the diffference is that my Unstable has a standard ext4 /home partition, whereas my Testing uses a LUKS-encrypted /home partition... identical to my real installations [Stable] on Tower & Lappy -- gulp.
You know, when updating, considering the limited nature of pacman.log, would it be "better practice" to pipe all terminal output into say a simple log file in directory home? That way, if another x crash occurs, we get a better picture of what updates are offered, any transaction interruption not logged by default, and which (if any) post hooks were run? For example:
pamac update -a |tee -a ~/pamac-updates.log (doesn't always work) sudo pacman -Syu |tee -a ~/pacman-updates.log
On my system i had the same "wait 2 minutes because of 'A stop job is running for User Manager for UID 1000' ” on reboot ... but to do a reboot i additionally had to do "reboot" in the terminal as my KDE installation did not show any entries anymore in the "KDE-Menu" except for favorites so the "Leave" (rather "Verlassen" in german on my installation) tab was empty so no option for reboot or shutdown.
Not quite sure i know how to answer this. Back [quite some time ago now, longer than i remember details clearly] when i first created this VM, i simply picked the offered option during the installation to encrypt my VM's separate /home partition with LUKS. So its disposition ever since then has been simply whatever/however the installer did it. Probably i have missed your actual point here?
You should have a systemd cryptsetup service for each encrypted partition, they derive their names from crypttab entries, you can list their exact names using ...
systemctl list-units systemd-cryptsetup*
The exact service name is also in your boot text grab.
If you only have one encrypted partition, thus only one crypttab entry, this command should return onw row.
For instance ...
UNIT LOAD ACTIVE SUB DESCRIPTION
systemd-cryptsetup@luks\x2d5834e487\x2d2a64\x2d44e7\x2dad6a\x2dad3e41509d5d.service loaded active exited Cryptography Setup for luks-5834e487-2a64-44e7-ad6a-ad3e41509d5d
Use that service name to plug into systemctl status command ...
systemctl stats systemd-cryptsetup@luks/...
Hopefully this give a meaningful error message to help.
Not sure how Calamares configures luks for only one partition, I've only ever fully system encrypted.
What is in ...
Also look for potential luks configuration entries in
Did you check in journalctl for errors during the actual update?
Tis such a PITA not being able to copy & paste any text from/to the VM's tty.
Oh, i see... but it confuses me. If you zoom in on that first pic [my earlier post i mean], there are some weird characters [incl. apparent rectangular blobs] that i have no idea how to reproduce when i laboriously type in the long command string you advised.
Ah, but WHERE exactly do i begin? With the first of those strange backslashes [& why are they BACKslashes anyway?], or with /dev/mapper/luks-b6163f1a-... or with luks-b6163f1a-...?
Nada & zilcho.
I did nothing more, or less, than i wrote in my OP.
I'm going to take a stab now at guessing the answer to my earlier "where to start?" question is luks-b6163f1a-..., based on the contents of /etc/crypttab to which you cleverly directed me. I'll be back soon with an update once i finish typing it all in & see what happens.
Edit: Hmmm, slightly anticlimactic [even once i realised you omitted a u ]:
Aaarrrghhh, i'm a numpty, that horrid long string with the backslashes IS the unit i need. OK, trying again now...
Groan, presumably this is Not Good [note, your initial / should in fact be a \]:
I intuit this might be telling me that the critical crypto service has gone to join the choir invisible, has gone to meet its maker, bereft of life it's pushing up the daisies, this is an ex-service, etc.