Skip to main content
Topic solved
This topic has been marked as solved and requires no further attention.
Topic: zaplocked installation recovery SELF-RESOLVED? (Read 1873 times) previous topic - next topic
0 Members and 7 Guests are viewing this topic.

Re: zaplocked installation recovery

Reply #15
That's not a stress test. That's a benchmark
I think I've made a few suggestions in this thread to try and give the OP a hand reviving their install.

When I see posts that have little merit I let it go most off the time. But when I see a post, part of which is both of little merit, and potentially harmful I comment.
I said you were right about the the backup. I said I didn't see the benefit of a stress test because there isn't a benefit and it just increases the wear to your drive. The only way to properly stress test a drive is to write and read the entire drive, mainly write. Many times.
And as already stated it proves nothing about a likely failure if it passes. Waste of time and shortens the drives life.

Your opinions are not my policy.  Have a nice day.
Cat Herders of Linux


Re: zaplocked installation recovery

Reply #17
Gee, where to begin? Thank you every one. It's late, I'm past dead, and I won't be 'doing' anything today but I have downloaded the

Code: [Select]
artix-community-qt-openrc-20220627-x86_64.iso

image.  The disk has 10k hours, SMART reports no issues, don't think it was hardware but for now the magnetic 4tb WD is sidelined and already wiped.

Provisioonally and just trying to cut ahead of more involved footwork, how about installing without formatting?  I presume thos would preserve all my user-installed apps as they were. The 'home' being on an off-system path all the look-and-feel is well protected. Subsequent updates might fix any loose ends.

BTW what's the difference other than date between the above iso image and

Code: [Select]
artix-plasma-openrc-20220123-x86_64.iso.part
  ?





Who, has loved us more?

Re: zaplocked installation recovery

Reply #18
6/27 is newer....  possibly a weekly update unless they released it as a the official iso.  the .part means it wasn't fully d/l

as reported by @####### If the HDD is suspect, sure, replace and bin it after getting any useful data, but the first post specifically mentioned a lock up and hard power off which could cause corruption by itself on a non-COW fs, not every time, just if you are unlucky.

I have been unlucky like that on ext4 once in the years of using ext4.  I did reinstall and all was fine.  But i chose to test it. 

10k hours.  that's an experienced 4gb hdd.  416 days of continuous use  or 1250 days (3.4 years) of use 8 hrs every day.  I bet you have had it for longer than 3 1/2 years?

Sorry for your hdd and your problem.
Cat Herders of Linux

Re: zaplocked installation recovery

Reply #19
Gee, where to begin? Thank you every one. It's late, I'm past dead, and I won't be 'doing' anything today but I have downloaded the

Code: [Select]
artix-community-qt-openrc-20220627-x86_64.iso

image.  The disk has 10k hours, SMART reports no issues, don't think it was hardware but for now the magnetic 4tb WD is sidelined and already wiped.

Provisioonally and just trying to cut ahead of more involved footwork, how about installing without formatting?  I presume thos would preserve all my user-installed apps as they were. The 'home' being on an off-system path all the look-and-feel is well protected. Subsequent updates might fix any loose ends.

I'm not fully following that.
But yes if /home is on a separate partition or drive then once re-mounted onto a fresh install both your 'look-and-feel' and the majority of your programs settings would be preserved. Only where you'd made system wide changes in /etc would the settings be lost unless copied back into place manually.  Also system program data under /var would be lost.

You've mention installing without formatting but also having wiped the drive, or at least a drive. That's the bit that confused me ?
But assuming you still have the intact root filesystem I wouldn't recommend trying to install an iso over it without formatting. Though it's not something I've actually ever tried.
I think it would be better to chroot in and reinstall any packages which show as broken by the paccheck command I showed earlier or just reinstall every package automatically,. The same page linked before explains how
https://wiki.archlinux.org/title/pacman/Tips_and_tricks#Reinstalling_all_packages 
Code: [Select]
pacman -Qqn | pacman -S -

Quote
BTW what's the difference other than date between the above iso image and

Code: [Select]
artix-plasma-openrc-20220123-x86_64.iso.part
  ?
.part mean an incomplete download. Either still downloading or interrupted.

Re: zaplocked installation recovery

Reply #20
Gee, where to begin? Thank you every one. It's late, I'm past dead, and I won't be 'doing' anything today but I have downloaded the

Code: [Select]
artix-community-qt-openrc-20220627-x86_64.iso

image.  The disk has 10k hours, SMART reports no issues, don't think it was hardware but for now the magnetic 4tb WD is sidelined and already wiped.

Provisioonally and just trying to cut ahead of more involved footwork, how about installing without formatting?  I presume thos would preserve all my user-installed apps as they were. The 'home' being on an off-system path all the look-and-feel is well protected. Subsequent updates might fix any loose ends.

BTW what's the difference other than date between the above iso image and

Code: [Select]
artix-plasma-openrc-20220123-x86_64.iso.part
  ?







so installing without formatting has been something i have done in the past with mixed success and failures.  IF you did that there's a chance that versions will be so out of whack dependency wise that you won't get a pacman -Syu to even run. 

I would honestly try what @gripped has said first.  It's a good solution and one i hadn't tried before but will def keep in mind should i ever have such a problem like that again.  If you still have the drive and haven't formatted the root directory that is.  chrooting in and just reinstalling the missing or damaged files is probably the best option.  It's not something i would have tried as i didn't know how to check for what's missing but @gripped has told you how to check for that.

there is no need to wipe /home or format it at all.  whatever you install root on you can keep /home just like it is now since it's on a different drive altogether.
Cat Herders of Linux

Re: zaplocked installation recovery

Reply #21

Code: [Select]
artix-community-qt-openrc-20220627-x86_64.iso

...
BTW what's the difference other than date between the above iso image and

Code: [Select]
artix-plasma-openrc-20220123-x86_64.iso.part
  ?


the *.part is my bad, I know what that means, the difference lies in that one of the iso's is near twice the size of the other, aren't both of them 'plasma'?


Quote
You've mention installing without formatting but also having wiped the drive, or at least a drive. That's the bit that confused me ?
{/quote]

Ok, I see where I messed up the comms :-)

The filesystem corruption causing data-loss which I subsequently recovered form was a 4tb DATA drive also hosting some of the home folders including my usual one..

The possible cause was a lockup of a Linux system that is on a 1tb ssd with some other distros. This filesystem event was internal to the bootable partition and is additional to and likely the cause of the other one affecting the entire 4tb data drive which took a fit upon journal replay following the lockup. I don't remember which system locked up, but my Artix system is the ONLY one with any teltale issues, the primary subject of this thread.


Who, has loved us more?

Re: zaplocked installation recovery

Reply #22
what's the difference between plasma and qt you ask...  plasma is traditional kde default setup.  qt probably has some other taskbar.  it's not lxqt so likely not lxde-ish but frankly i have never used either..  i could.  i d/l them all via qbitorrent and could easily pop that in a Vbox and see..

plasma is half the size f qt.   would guess thats the eco version for lightweight setups


Cat Herders of Linux

Re: zaplocked installation recovery SELF-RESOLVED?

Reply #23

10k hours.  that's an experienced 4gb hdd.  416 days of continuous use  or 1250 days (3.4 years) of use 8 hrs every day.  I bet you have had it for longer than 3 1/2 years?
It's all part of the package :-)
Quote
Sorry for your hdd and your problem.
The last disk I sent to hard-drive Arlington was a 2tb also a black WD with, if I remeber correctly, over 30,000 hours on it. 


BTW just minutes before booting Artix succesfully I had a lockup on a Devuan system, again with a journal replay on the DATA disk (brand new 2tb) with 'home' on it upon reboot also to Devuan. So I have two chicken-and-egg questions in mind:

1
Does a lockup cause a fs/disk problem resulting in a journal-replay or is the lockuop the result of the fs/disk problem? 

2
Does a reboot to a DIFFERENT distro following a lockup cause a journal-replay or is that journal-replay the potential cause of other fs problems, this time in the booted system partition? 


Who, has loved us more?


Re: zaplocked installation recovery SELF-RESOLVED?

Reply #25
you using some good firewalls and some vpn there hossenfeffer?  Something def wonky going on there.
Cat Herders of Linux

Re: zaplocked installation recovery SELF-RESOLVED?

Reply #26
I got nearly 19000 hours on my current ssd, I usually just get cheap used ones, rarely get any trouble. With electronics new things have a higher failure rate, then it reduces until they start to hit their typical natural lifespan whatever that is.
 Re the questions, usually the only way to determine what is going on is testing, you can look at SMART logs and also at your syslog to see if there are any relevant errors. It's not impossible (just unlikely) there could be some mobo hdd controller fault too. There have been software bugs that cause disk problems too, but that would be extremely unlikely to hit any release software. I suppose it might affect other partions, one could vaguely imagine if something got written in the wrong place. Fix all the software problems and if they keep coming back then you have a hw fault I guess! I usually wipe and test drives with badblocks -ws before use, it can take a long while to run, hours or days on a big one, although with ssd's you use some other method to zero them. I have seen suggestions to leave some free space at the start and end of the drive when partitioning to allow the drive to internally manage any faults.


Re: zaplocked installation recovery SELF-RESOLVED?

Reply #28
you using some good firewalls and some vpn there hossenfeffer?  Something def wonky going on there.


who, what, huh?????

while I'm here again

1
I'm NOT getting emails when new posts arrive so I end up looking like a loser not interested in the replies

2
Still getting (much less frequent) loclks/journal-replays but no corruption. The 4tb iron is wiped and ready to be given as a gift to someone I don'yt like.

3
I can duplicate one of the issues raised and it has nothing to do with the topic,

"/sbin/init does not exist"

it's deservoing of a new thread which sjhhopuild also reduce the confusion in this one :-)

  
Who, has loved us more?

Re: zaplocked installation recovery

Reply #29
The OP didn't state how the corruption occurred. It may be a bad disk. It may have been a bad shutdown, power cut. IDK.

Yes I did, a system lockup forcing a hard reset. I would like to hope that all Linux systems always get ready for the power-failure about to strike, maybe they do maybe they don't. A lockup could of course visit even during such a wise process, the fact is I don't know what caused the lockup.

So following this lockup/hard-reset the DATA drive went into extensive journal-replay the result of which was possibly a directory corruption on the data drive. This was ONE problem, subsequently fixed.

The OTHER problem involved Artix, and it was/remains less clear.  The 1st part of this one was missing kernel and initrd entries under /boot, this was fixed by a chroot and and a system update which fortunately 'included' a kernel. The second part of this same problem was a FALSE-ALARM.  I had seen it before but this time, cought up in the turbulent situation, I just fixated on a new expression I had not seen before in the Artix error message (the one about initd) and didn't see the one about /dev/sdb right next to it with which I was already familiar.  So it SEEMS like this second part of the second problem had nothing to to do with any data-drive or Artix system corruption.

 And thanks for your help, needless to say, everyone else's too.

    

 
Who, has loved us more?