Hi, just checking in to add that I had the same behavior after a failed pacman -Syu update.
Both PamacGUI and pamac update or pamac upgrade did freeze the machine after eating the whole RAM in 10 seconds or so.
Also, I had pamac checkupdates already fired just after the session login, so the PC started to freeze over after 4/5 seconds or so after logging in. To remove it I just went to Pamac settings ( “Add/Remove software” → “Settings” → Unticked “Check for updates”).
As for pamac update/upgrade the process got stuck at Synchronizing package databases... : after that I terminated it before the machine froze again.
As I mentioned, this happened after a failed attempt to update. I can not replicate the error I had but I can remember multiple lines of
...
error: could not read db 'extra' (Damaged tar archive)
error: could not read db 'extra' (Damaged tar archive)
error: could not read db 'extra' (Damaged tar archive)
...
While this happened, pacman worked perfectly fine:
[nicola@nicola-desktop ~]$ sudo pacman -Syu
[sudo] password for nicola:
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
multilib is up to date
:: Starting full system upgrade...
there is nothing to do
[nicola@nicola-desktop ~]$ sudo pacman -Syyu
:: Synchronizing package databases...
core 150,4 KiB 1419 KiB/s 00:00 [################################################] 100%
extra 8,7 MiB 10,5 MiB/s 00:01 [################################################] 100%
community 29,0 B 1705 B/s 00:00 [################################################] 100%
multilib 143,9 KiB 3,80 MiB/s 00:00 [################################################] 100%
:: Starting full system upgrade...
there is nothing to do
[nicola@nicola-desktop ~]$
I’m writing as past tense here because, as I’m tinkering, the problem seem to have disappeared by itself!
I’m just writing to resume and signal a problem (bug?) already signaled (I looked over the forum to post to already opened threads but I could not find one open).
The problem is that when trying to update from Pamac GUI or pamac update the process starts to eat up all the ram. Then after 30 seconds or so it goes back to normal as the pamac process crashes (I think it crashes, at least). Then again after 10 minutes the process repeats.
Here’s a photo (I had to take a picture with my phone because the os was completely stuck):
(I stopped it before my machine would freeze again). While this happens with pamac, pacman seems to work fine:
[nicola@nicola-desktop ~]$ sudo pacman -Syu
[sudo] password for nicola:
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
multilib is up to date
:: Starting full system upgrade...
there is nothing to do
I had this behavior for the first time after a failed update: my internet connection got lost and pacman -Syu got stuck. After closing the process and running the command again I got multiple errors like this:
...
error: could not read db 'extra' (Damaged tar archive)
error: could not read db 'extra' (Damaged tar archive)
error: could not read db 'extra' (Damaged tar archive)
I’ll just add that these issues are possibly a result of not maintaining your system adequately. This isn’t a random comment without merit: The community repository has not existed since mid-late 2023 (several months) – and yet, it hasn’t been removed from your configuration.
It’s a simple fix (a manual edit), but that’s not the point. It indicates a broader problem with not observing the manual intervention notes as they appear in the Stable Update Announcements - and there have been many in recent months; especially .pacnew file instances that should have been merged (by you) with respective configuration files in your system.
Please see Check and manage pacnew files for a brief introduction to these needed modifications, and perform some additional research when possible.
I really thank you for your suggestion. I apologize if my system may look not well maintained, but I never had the chance to dive deep into this kind of concepts. So when a problem comes up, like the one you are raising above, I may not even have recognized it as a problem, considering that the system still worked perfectly until yesterday.
Anyway, I’ll make sure to pay more attention in the future and read along the sources you linked asap.
Thank you again for letting me notice this!
Attention is one thing; more important is to forge a regular habit of checking for issues. Manjaro is a rolling release and as such we are expected to monitor the status of updates as they are announced. When each update is ready (usually two or three per month) the Update Announcement is released.
A good recommendation is to wait a few days (or a week) when this happens and see what problems (and fixes) are discovered and listed under the announcement. This gives a certain advantage, rather than blindly running sudo pacman -Syu with your fingers crossed; at the mercy of dumb luck.
Not everyone has the aptitude, the time or the patience to maintain Linux on a rolling release cycle; as all Arch based distributions are. Manjaro is a great OS, but you really must take care of it. For those who prefer less frequent updates and generally more stability, for longer periods of time, we often recommend moving to a distribution with a point based release cycle.
Don’t let this deter you, though;
Manjaro just takes some determination.
Theres nothing blind or lucky about it.
You passed specific arguments to a specific tool that should commit certain actions.
Once entered you will then be asked for admin password and presented with the proposed changes.
You would have to be literally physically blind for this to ‘blind’.
PS.
Also of note is that this advice of ‘waiting’, intentionally or otherwise, skirts the issue of security.
We have a pretty salient example from yesterday with xz.
Depending on the current sync status of a system, following said advice could result in a prolonged period of vulnerability.
Some people do seem to consistently struggle with nuances of forum commentary. However, @Mirdarthos has again captured the essence of the intended meaning, through interpretation; and I therefore award him the unofficial be-bop emoji.
Blindly performing an update without having made certain there are no outstanding issues that might prevent its completion, is not recommended.
Without taking that ounce of prevention into consideration before an upgrade, error-free completion can only be attributed to dumb luck.
If you want to fundamentally change some meanings, and include self-defeating statements, I guess I will accept that as the intention of your statement.
Still … in this example, at best we have then equating ‘waiting’ with being ‘not blind’.
Which, I would hope is self-evident enough, but in case it is not … how does an otherwise ignorant user who cant/wont inspect or understand the package changes become any more informed simply by waiting a few days?
Ah well … semantics again maybe?
I would suggest folks become familiar enough with the system they are administering to not be blind.
Right, so, the ‘not blind’ user is someone who is continually ignorant but waits a few days seeing what bubbles in the release announcement?
I think I disagree. And its rather poor advice in the context of, for example, critical security patches.
But alas maybe there is nuance again and of course I should have understood that waiting a few days in fact does not mean waiting a few days if there are security concerns … beep boop, none of this is any longer pertinent to the thread either.
I’m trying to resolve a system instability. Whenever pamac checkupdates runs, it quickly uses all of my RAM and swap and then the whole Manjaro operating system freezes and has to be hard reset at the desktop.
I’ve recently become aware of the closing of the community repository through stable update 2023-07-10 which needs to be addressed.
And I’ve tried the command line fixes from the second post of the following forum post with no fixing it.
running pamac checkupdates results in more errors than I can show.
