I hope the title is exaggerated, but that’s how it’s looking at the moment.
Summary:
Laptop [TW KDE 20170629 or 30, can’t remember now] suddenly began not responding to YaST/Zypper about 90’ - 120’ ago [maybe ~11:00 - 12:30 UTC], during a lengthy bout of me manually installing multiple software [assorted One-Click & RPMs]
System Monitor showed Btrfs using 25% cpu continuously once #1 became noticeable to me.
YaST couldn’t be killed
Decided to try a logout, but command [via Application Menu] was ignored.
Tried a reboot, but ditto.
Tried to get to tty2, but ignored
Forced a reboot.
Since then, cannot get back into TW. Every boot attempt results in a black screen of complex text, or occasionally i do get the boot menu, but then choosing a normal boot option leads to either more text, or sometimes to the Home partition unlock screen [root is Btrfs & NOT encrypted, Home is ext4 & Encrypted [from installation]], but after inserting my password i get stuck indefinitely at the TW splash screen with the 3 horizontal dots scrolling back & forth continuously.
A couple of times i then tried to use the Snapper rollback capability from boot [have never used it before, & am unsure how to drive it]. It baffles me by showing timestamps completely unrelated to my local timezone - does it use maybe UTC instead of my local zone = Australia]? Pursuing my guess that it’s UTC, i selected an entry lying near the start of the time range i mentioned at #1… but i don’t yet understand what to do next… am about to research that now, but want to post this initial cry for help as a placeholder.
This possible catastrophe is very upsetting given [per my other recent posts] i am new to openSUSE & TW, but have been intensively testing TW on my reformatted Lappy [the now unwell pc as above] for a couple of weeks & had been super-impressed with it… tomorrow i had planned to then also convert my Tower to TW… but that’s now completely up in the air.
Hmmm, lots of websites tell me how wonderful Btrfs, Snapper & Rollbacks are, but either do not give step by step sequences of the actions a desperate & confused user needs to do at boot to benefit from this capability… OR… do give some steps but which don’t correlate with what i have on my screen.
Taking a guess, i scrolled down the list of pre- & post-snapshots, picked one that might have been near the beginning of this mess, pressed enter, saw a new screen of 3 entries with cursor highlighting middle one. Before i could finish reading & trying to understand each one, pc self-entered that highlighted row, which seems just to be the standard boot option “openSUSE Tumbleweed”, & the sequence i described in first post continued, ie, no progress. I forced a shutdown then after a while booted again, this time picking the same snapshot but ensuring i then quickly moved the cursor to select the 1st row on the next screen, ie, “Bootable snapshot #4xx” [maybe #410; it’s now ~ 01:11 (or 15:11 UTC) & i need sleep, so details are beginning to elude me]. Pressing Enter gave a new single row screen, “If OK, run ‘snapper rollback’ & reboot”. Huh? Where do i run it, how do i run it? Pressing Enter on this row merely briefly gives a black screen then the same screen reappears, etc, etc.
Wondering if maybe that meant the snapshot was damaged, i escaped back out to the top level & scrolled down the snapshots list & picked a much earlier one, which * should be a couple of hours before the initial symptoms as described in my Summary [but i had been installing software on & off all day, so who knows when the fault really began?]. This one was “Bootable snapshot #395”, whose description was “openSUSE Tumbleweed (4.11.7-1,2017-07-03T09:29,pre,yast sw_single)”. However, all the same behaviour repeated as i described in my previous paragraph with the later snapshot.
Clearly i have no idea how to use these snapshots.Maybe after some sleep it might make more sense to me.*
Thanks for replying. I need to amend part of my initial post by inserting more info, as your remark made me realise i gave you the wrong impression by omission. Hence:
Laptop [TW KDE 20170629 or 30, can’t remember now] suddenly began not responding to YaST/Zypper about 90’ - 120’ ago [maybe ~11:00 - 12:30 UTC], during a lengthy bout of me manually installing multiple software [assorted One-Click & RPMs]
System Monitor showed Btrfs using 25% cpu continuously once #1 became noticeable to me. I left it alone until that process disappeared from the top of the list [maybe an hour, maybe more], whence the next highest & unrelated] process using cpu was only 1 - 3%. However still nothing related to YaST or Zypper would respond to me
. 1. YaST couldn’t be killed
Stuff like psensor, recoll, netactview & much more], all from Search, & all by using the TW ones. “Not a good thing”; are TW users not supposed to install any standard software?
Well gosh, yes, but that’s a bit harsh… after all Richard Brown puts out YouTubes etc encouraging adoption of TW because of the safety net of Btrfs + Snapper… which is exactly why i became interested in TW as a potential candidate for me.
About snapshots… and yes, relying only on the timestamp might be misleading or non-informative…
Depending on the age of your machine, you can easily assume that earliest snapshots were created soon after initial install.
You should know when snapshots are created…
Snapshots are created daily by default, but in most cases you’re probably most interested in snapshots created within the past 24 hrs or so.
Snapshots are created before and after every software install (using zypper or YaST).
So, if you can keep a record of when and what you’re installing each time, you should have an easy idea what your system’s file status was recently. Count pairs of snapshots backtracking from where you are now, or just roll back to the beginning of the day to undo everything you’ve done that day.
A basic concept is that snapshots always move forward.
That means that whenever you roll back, you’re not undoing your snapshot history, you’re merely creating a new snapshot reflecting the rollback. This also means that since a rollback doesn’t destroy, you can decide to undo or rollback to a different point in time… But keep track of what you’re doing because if you rollback to multiple points in time this will become very confusing. A better practice is to determine where you can be sure you can restart with a clean slate and rollback to that point instead of trying for a more precise point in time.
Lastly,
I think you’ll be happier using the Snapper CLI instead of YaST unless you’re looking for something that YaST provides exactly.
The Snapper CLI can display much, much more information, specific types of information and execute faster than YaST.
Richard also always points out that users should not add home: and devel: repos. If you want/need a package in TW, ask the maintainer to push it to TW. If the quality is good, and TW with it still passes openQA, you know for sure it’s not built against the wrong libs etc etc.
Hi
So psensor and netactview from my repo (It’s not standard software)?
Don’t use one-click or YaST for software and updating in tumbleweed, use zypper from the command line so you can see exactly what is happening, zypper should also be used for upgrading after every new release…
Download the rpm’s and create a local rpm repo, or use zypper in <url_to_rpm> likewise add verbosity with -vvv so you can see what the system is going to do.
For recoll, have no idea, that’s from a KDE repo… what else it may pull in (or state/compatibility) when adding.
What a relief - i finally found a Snapshot that was not damaged, & thus have managed to roll-back to it, then perform the usual
sudo zypper dup --no-allow-vendor-change
to become TW 20170703 [it was on 20170630 when last night’s mess occurred]. All seems ok now, phew, so maybe tomorrow i’ll have time to dive back into this thread & try the actions you suggested / answer your questions. I’m tired but grateful for your assistance… albeit fervently hoping for no more “excitement” like last night…
I did not read all of this thread, but I wonder about your excitement. At the most you can be forced to re-install Tumbleweed and restore your data from your backup. Which will take much less time then spend on this thread. Not that I want you to not start a thread like this. Such a thread is always good for learning and understanding for you as well as others. But I do not understand the panic.
As i said, they bore no superficially obvious relationship to the times i was actually working on my Lappy, in terms of my local AEST domestic timezone. However, when i assumed they were in UTC & then did the necessary arithmetic, they began making plausible timing sense. Rather than me merely assuming though, it would be nice for me to know for certain that they are UTC [or not, as the case may be].
Well, now that i have been so badly bitten, in future i’ll make an effort to note times as you suggested. But just fyi & to provide [self-defence] context, i am an openSUSE newbie, & indeed have only been a fulltime Linux user since early 2014. Probably upwards of 95% of my Linux “experience” [such as it is] is wrt Ubuntu/Debian-based distros, & thus using apt in cli & Synaptic in GUI. I’d never even heard of zypper & YaST prior to deciding within the past 2 months to have a look at openSUSE, & the amazingly complex multiple repositories with clashing package versions &/or dependencies, has knocked me sideways – it’s an utterly new paradigm for me, & one that i am clearly not yet adept with. In my Ubuntu/Debian-based distros paradigm the repos were substantially easier for me to understand & work with, i had no comparable trouble discerning “danger” from “safety”, & i had no need for careful note-taking in terms of times that i installed stuff. Thus, to the extent that in openSUSE i might well be the architect of my own distress, is directly caused by my lack of comfort yet with this paradigm & its conventions. If i decide to stick with openSUSE, i shall certainly be trying to improve myself.
This was very helpful - thank you. Indeed, the night after that distressing night of my troubles, when i posted that i’d got my Lappy going again by finding a good Snapshot at last, came about because i specifically applied your “determine where you can be sure you can restart with a clean slate and rollback to that point instead of trying for a more precise point in time” logic.
These were good to read thanks. FYI, this one ended up being especially helpful in allowing me to better comprehend how i was supposed to use Snapshots for actual rollback… https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.reference/cha.snapper.html#sec.snapper.snapshot-boot [Section 3.3]. What had totally confused me with my initial many attempts to rollback was, as i’d posted at the time, the way that no matter what snapshot i’d tried at boot, each time i Entered with it selected, the screen went black then “bounced back” to the same screen again… & if i tried again, it bounced again… & this happened for many snapshots that i’d tried. This was distressing as i did not know if it meant that i was not doing the procedure correctly, or if all those snapshots were bad & if so, HOW/WHY?], or if all the boasts about how good Btrfs & Snapper were were… false. Once i eventually, the following night, found that “good” snapshot, using the identical method as previously, i deduced that simply by bad luck or whatever all the other snapshots i had tried, were corrupted, but my method logically had to be valid or else my Lappy would still be dead. I know that everything i write here will seem cringeworthy to all you long-experienced openSUSE wizards, but to a distressed newbie it was confounding.
Not sure if you’re alluding to getting into cli at boot, or after booting [which was impossible given the whole issue was inability to boot], but i did look at it once i let the onscreen failed attempts to decrypt Home timeout & fall back to cli in “emergency mode” *. As you said, there were copious options available [they scrolled off the top of the screen & i could not see how to scroll back up to them]. However i did then, in cli, try multiple snapshots, but every time the response instantly was:
Pls note my blurb in my earlier reply to tsu2, for context.
want/need a package in TW, ask the maintainer to push it to TW
Realistically, why would any Dev take any notice of me? Also, if they did take notice of me & then push it across, couldn’t they have already realised to do that on their own volition… i mean, it’s TW after all… wouldn’t it be self-evident that TW should have their stuff, without a small-fry random user needing to prod them?
As a result of various once-off programs [that i rely on in Mint & Maui] for which i had to use One-Clicks [or otherwise the RPMs thereat], my YaST now has quite a few non-standard repos as well as the standard ones. However, i am aware of the potential problems consequent of that, & so as soon as said pgms are installed i disable those repos. As a result, my Enabled repos list is only:
Thank you Malcolm. Pls note my blurb in my earlier reply to tsu2, for context. If that’s the necessary program / package installation paradigm, then i suspect more & more that i made an error of judgement in coming to Tumbleweed, or maybe even openSUSE overall. I just don’t think i’m good enough to make the cut. It’s a pity, because TW KDE has given me THE very best Plasma experience i’ve ever had, of any distro i’ve used or even just tried since late 2013. However, the management & maintenance of TW re repos & package installation looks like being beyond my capability / desire.
Realistically? Why wouldn’t they? I’ve seen it happen dozens of times: post on the opensuse-factory@opensuse.org mailing list, ask for addition of a package to the distribution. The package will be reviewed and if it’s stable and well-maintained it will be pushed to TW. It’s the way it works, real reality :D. We should get rid of the idea that no dev or packager is interested in the user. Me myself f.e. loves some praise of users after fixing/improving something.
As a result of various once-off programs [that i rely on in Mint & Maui] for which i had to use One-Clicks [or otherwise the RPMs thereat], my YaST now has quite a few non-standard repos as well as the standard ones. However, i am aware of the potential problems consequent of that, & so as soon as said pgms are installed i disable those repos. As a result, my Enabled repos list is only:
Visible repos shouldn’t be an issue. What can be an issue is that some packages from these disabled repos might, by dependency, block “normal” updating procedures. Suppose you installed package X-2.1 depending on the distro’s libY0-2.1 . X-2.1 is the only version known to the system, repo is disabled. Now TW moves to the new libY2. On your system it won’t since X-2.1 depends on the older version. By asking for submitting X to TW, the packagers/devs have the opportunity to see if X builds against libY2, and when they’ve made it to do so, tested it, they push it to TW and *you *can use it without having to break your TW.
Pls note my blurb in my earlier reply to tsu2, & indeed also the other respondents, for more context.
In addition to my various earlier words, & to try to explain to you why that drama was a drama, & quite stressful for me… as a consequence of my inexperience with openSUSE, & [as my forum name might rather indicate] my non-openSUSE Linux “experience” being inclined substantially more to GUIs than CLI, i have found myself frequently “off-balance” in my openSUSE explorations & workings. There’s still so very much i don’t know, don’t understand [despite my days & weeks of research & desperately trying to learn; i’m much better than i was a month ago, but still hopeless if measured against all you experts]. The repos continue to puzzle & frankly scare me; i’m simply not yet good enough to grasp their vastly greater complexity than i’m used to with Ubuntu/Debian-based distros like Neon, Mint & Maui etc [where, if i try to install a package via [i]Synaptic, or apt in cli, it routinely “just works”, or at worst & only rarely] throws up a missing dependency i need to track down. In the worst possible case the desired program won’t run, & so i merely remove it again & move on. Never before have i experienced the drama of actually destroying* my OS merely by trying to install a commonplace program. Hence the shock i received when this recently occurred, played a large role in my emotion-imbued post initiating & conducting this thread.
*of course, later i came to know that was not the case, but at the time the prognosis looked grim to me.
Furthermore, I’d spent what felt like a long week setting up my Lappy in TW to be “just right” in terms of my personalisations… resolutely tracking down all the Mint/Maui programs i need to have in my TW, & sometimes feeling i was fighting all the way not only to even find many of them, but to install them successfully. Yes of course i knew on that Bad Night that i could always “just” start again, but the prospect of then having to also start all over again getting all those programs depressed me.
Additionally, i’d also extensively personalised my Plasma Desktops & Activities, & where appropriate many individual programs’ settings too, & was very unhappy about possibly losing all of that work too. Yes, i do know that all those config data reside in my [separate] /home partition not the / partition, thus shouldn’t be lost if i needed to reformat / to reinstall TW. However & here i confess that my sleep-deprivation had partly clouded my judgement] given that i chose to encrypt my /home & swap partitions during installation in mid-June, i was very scared that if root was corrupted [as i had assumed to be the case give the pc would no longer boot, as posted] & had to be reformatted, that maybe that might also then render my /home permanently impossible to decrypt? I suppose you’re now rolling around laughing at my ignorance, but this is yet another aspect of oS TW that i still don’t understand compared to *buntus - eg, in Mint & Maui, /home encryption during installation creates an .ecryptfs directory in User’s /home, NOT in root, thus ongoing capability to correctly decrypt /home is independent of whatever travails might befall root. In contrast, in my oS TW file structure i cannot see evidence within /home of some oS equivalent to .ecryptfs, & thus… late on the night when the Trouble erupted, after a long week & with bad sleep deprivation, i did seriously fear that a possible need to blow away root might mean also blowing away my encrypted /home.
Finally, & extending from that previous sentence, during the day of the night when the Trouble occurred, i had spent many many hours resolving a longstanding problem with my Win10 VM [earlier this year, weirdly, it decided to Inactivate itself, & i frankly couldn’t be bothered to waste the hours i imagined would be needed to deal with MS trying to persuade them that it is legitimate & i am not a pirate]. On that recent day however, my oS TW was really looking & running great, after the week i’d spent working on it as described, & so i decided now the time was right to also resolve that Win10 Activation [successfully]. During that same session the recent major upgrade also occurred, after which i took the time going back deep inside the new version to once again turn off many of its inbuilt privacy-stealing settings, & then cosmetically personalising its desktop etc. This VM resides in /home… hence per the preceding para, i was mortified that IF i lost root & then /home, i’d also lose my VM.
I fear that possibly you still “do not understand the panic”. If so, the only other tactic i could try is this… pls cast your mind back a long time ago… there surely was some point in your past when you also knew as little about Linux in general & oS TW specifically, as applies to me currently…?
Don’t panic. Many users have converted all their machines to Tumbleweed, e.g. openSUSE Tumbleweed: A Linux distribution on the leading edge | ZDNET I made the transition last autumn, when I realized that the software is stable and up to date. Now all my computers have Tumbleweed installed. zypper dup --no-allow-vendor-change never failed on any of them. Updating from 4.11.1 to 4.11.8 / 20170703 changed some 2600 packages without a single glitch:
Jul 04 15:35:43 hofkirchen systemd-journald[28582]: Journal stopped
...
Jul 04 15:35:55 hofkirchen kernel: Linux version 4.11.8-1-default
...
Jul 04 15:36:13 hofkirchen sddm[5201]: Session started
Ooooo, don’t underestimate others: I had just started my own business, was spreading cards etc, when I accidently borked the server that held my brandnew website etc. Don’t remember what I did but IIRC I installed some openSUSE rpms on it that rendered it unbootable.
To simplify openSUSE:
don’t add (packages from) home: and devel: repos. Things in there are breaking all the time because it’s the place where they should break
System stuff? Use YaST. If you can’t find what you’re looking for, use the software manager and search for ‘yast2-’, but, YaST all the way.
be aware that some packages you know from *buntu (derivates) may have (a tad) different names, or great alternatives. I sometimes use Mint and use other apps than on openSUSE to achieve the same things. One gets used to that too
don’t compare openSUSE to whatever, take it as is, learn and appreciate.
Re. posting here:
Even though I feel with you in re. ‘drama’: please, make up a better title than this one. My initial reflex was to skip it
Try to be brief, these long posts take a lot of time to process
If you refer to things you wrote before or elsewhere, please provide a link.
Never panic, we’re here
And most of all:
Enjoy and have fun finding your way around.
Know that there’s a friendly community that loves to help you aquiring this
On Wed 05 Jul 2017 11:26:01 AM CDT, GooeyGirl wrote:
malcolmlewis;2828662 Wrote:
> Hi
> So psensor and netactview from my repo (It’s not standard software)?
>
> Don’t use one-click or YaST for software and updating in tumbleweed,
> use zypper from the command line so you can see exactly what is
> happening, zypper should also be used for upgrading after every new
> release…
>
> Download the rpm’s and create a local rpm repo, or use zypper in
> <url_to_rpm> likewise add verbosity with -vvv so you can see what the
> system is going to do.
>
> For recoll, have no idea, that’s from a KDE repo… what else it may
> pull in (or state/compatibility) when adding.
Thank you Malcolm. Pls note my blurb in my earlier reply to tsu2, for
context. If that’s the necessary program / package installation
paradigm, then i suspect more & more that i made an error of judgement
in coming to Tumbleweed, or maybe even openSUSE overall. I just don’t
think i’m good enough to make the cut. It’s a pity, because TW KDE has
given me THE -very best- Plasma experience i’ve ever had, of -any-
distro i’ve used or even just tried since late 2013. However, the
management & maintenance of TW re repos & package installation looks
like being beyond my capability / desire.
Hi
Well OBS can be daunting with it’s package availability Also your end
use, do you really need the latest and greatest to do what your
wanting (features)? To many it’s not just the desktop but the packages
to do the tasks your wanting and compromises may be required.
Why not look at Leap then?
–
Cheers Malcolm °¿° SUSE Knowledge Partner (Linux Counter #276890)
openSUSE Leap 42.2|GNOME 3.20.2|4.4.73-18.17-default
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below… Thanks!
On Wed, 05 Jul 2017 13:36:02 +0000, GooeyGirl wrote:
> KARLMISTELBERGER & *KNURPHT Thank you both for your wise words, &
> your encouragement.
>
> PS - Re “*Don’t panic” … haha, am i over-thinking it if i were to
> intuit a HHGTTG inference? rotfl!
Not at all - there are quite a few fans of H2G2 in the openSUSE community
indeed, the major version number for Leap is a hat-tip towards it (and
openSUSE’s legacy of using 4 and 2 together for various versions early in
its history (0.42 for YaST IIRC, 4.2 for the original release, and so on).
Always know where your towel is. Mine’s on a shelf behind me (seriously).
Also your end