On Wed, Oct 06, 2021 at 10:44:15AM -0400, Felix Miata wrote:
Steven D'Aprano composed on 2021-10-06 21:03
(UTC+1100):
Thank you to everyone for your suggestions, I was
especially happy with
the look and feel of MX Linux. I'm used to booting up with Fedora taking
five or six minutes (going up to 20 or 30 minutes when it runs a file
system check)
That never happened to me using Fedora whether with Plasma or TDE.
I'm happy for you that your experience has been different.
- Fedora's
file system check on a failed SSD is crap. 22 hours of
checking, to get nowhere.
Fedora's fsck is the same one other distros use. A failed SSD will be a problem no
matter what is on it.
A file system check is not just fsck, it is also the user interface and
the reporting of errors to the user. Or in this case, the *lack* of
reporting of errors to the user. The only feedback you get is something
close to:
Running start task File System Check something something /dev/something/long string of
random alphanumeric characters
at the boot screen. You certainly don't get any feedback that the check
was failing with thousands of I/O errors. To see that, you have to boot
into the machine and look at dmesg, which you can't do until the fsck
completes, which it won't do until there are no more errors...
(One wonders if anyone at Red Hat or Fedora checks their boot scripts
on a machine with a failed HDD or SSD. I suspect not.)
With no feedback as to why the system was unable to finish, it took me
literally two days to work out that it was a dead SSD, by which time it
was too late in the weekend to purchase a replacement which meant I lost
another day.
(I know, I know, we're living in a time when four or five million people
have died due to the worst industrial accident in history, and tens of
millions more have lost their jobs and livelihoods. I should count my
blessings. What's three or four extra days to repair the system? Well,
in my case, aside from the frustration, it's about a day or so of lost
work.)
I eventually worked out what was going on by hard powering off the PC,
booting into a rescue disk, identifying the failing partition by its
UUID (it was /boot) and commenting it out of fstab. That allowed the
system to boot up fine from boot-efi. And that allowed me to fail the
ssd from the RAID array, and get back to work for a brief time, except
that I didn't realise that swap was also raided, and boy oh boy is it a
bad idea to be swapping memory to a failed disk.
Go on, ask me why swap was raided.
- Never, ever
enable "Fast boot" in the BIOS.
It works OK here on the few PCs I set it on, mostly Dells.
On my PC, in fast boot mode, keyboard and mouse is disabled until after
the BIOS passes control to the OS, which means you can't hit F2 to get
back into the BIOS to disable fast boot mode or change the boot order.
Fedora is considered by many to be "bleeding
edge", which sounds like the
opposite of your interest: stability.
The Fedora community doesn't consider themselves to be bleeding edge,
but they sure as hell behave like they are, if not right on the edge,
close enough to the edge to be splattered.
And you are right about stability. If I could, I would still be running
KDE and Linux from 2005 or so. Aside from the web browser, and perhaps a
few video codecs, everything I use worked just fine back in 2005 and the
last 15+ years has just made the overall experience worse.
(Gods, I'm going to be yelling "Get off my lawn" and shaking my fist at
clouds next...)
Two days' downtime isn't normal for installing
an OS
I know. But I have learned to expect the worst, always, in any and every
upgrade. I expect that it will always take longer than expected, be more
difficult than expected, will break things that I didn't expect to
break, and invariably I'll end up with applications with a shittier UI.
I've so rarely been pleasantly surprised by upgrades.
I mostly avoid live distros, with main exception being
Knoppix.
Now there's a name from the past. I remember when Knoppix was gee-whiz
new technology. Imagine being able to run Linux from a CD!
--
Steve