News for hackers. Stuff that matters.
Re:Skipping fsck during boot with systemd? by atoponce on Sunday December 21, 2014 @07:47AM (#8779614)
Offline filesystem integrety checks should not be stopped.There is a reason they take place to begin with. From the OP, it sounds like Eduardo takes every opportunity to stop filesystems checks from happening when booting. So long as he has a policy of doing them manually, and understand why they take place, that's fine. So, instead, he should just:

# tune2fs -c 0 /dev/sda1

Or whatever his filesystems are. Then again, this is risky, and the OP should understand why.

Re:Skipping fsck during boot with systemd? by boomlinde on Sunday December 21, 2014 @08:57AM (#8779749)
I don't that there is any basis for your assumptions about the poster. What makes you think that he takes every opportunity to stop filesystems checks? Did you read a part of the message that I did not?
Re:Skipping fsck during boot with systemd? by atoponce on Sunday December 21, 2014 @09:03AM (#8779761)
His opening sentence makes it pretty clear that he's executed Ctrl-C enough that he's familiar with it, likely executed it more than once, and expected it to work in Systemd.
Re:Skipping fsck during boot with systemd? by boomlinde on Sunday December 21, 2014 @12:44PM (#8780307)
OK, so it's pretty clear that he's likely to have pressed ^C more than once. That sure is a strong foundation for saying that it sounds like he takes every opportunity to stop filesystems checks.

... Especially since he didn't go out of his way to explain the exceptional circumstances that prompted him to try to skip fsck in the reported case... or did he?

Re:Skipping fsck during boot with systemd? by foxhill on Sunday December 21, 2014 @08:30AM (#8779702)
so your fix involves time travel, and you think that's reasonable?
Re:Skipping fsck during boot with systemd? by atoponce on Sunday December 21, 2014 @08:32AM (#8779709)
Time travel? Huh? I'm only stating that if the OP does not want his filesystem checks taking place on boot, rather than Ctrl-C them on every boot, to just disable them completely. Who said anything about time travel?
Re:Skipping fsck during boot with systemd? by foxhill on Sunday December 21, 2014 @08:36AM (#8779720)
his issue is not that fscks happen. his issue is that he cannot safely stop them when they happen.

so, you've given some information about setting the fsck counter to 0, so that it never happens. the only way anyone in this situation could utilise that information, would be to go back in time, run this command, and then boot again.

Re:Skipping fsck during boot with systemd? by atoponce on Sunday December 21, 2014 @08:46AM (#8779732)
No. You missed it. The OP stops filesystem checks happening fairly regularly. The great thing about computers, is identifying when you are repeating a behavior, to either have the computer automate your behavior for you, or to change your behavior entirely.

I am fully aware that the OP wants to cancel the fsck(8) in progress, and I am aware of both of the Debian and Systemd bugs in place to address the issue. Arguing on the Internet about whether or not he can or should Ctrl-C an active check, certainly when bugs already exist for both projects, is pointless.

Instead, to prevent additional filesystem checks, there are tools to make this possible. Unless he likes getting frustrated that he cannot press Ctrl-C during a check on boot, which doesn't seem to be the case. His system is booted now, so he has some options available to him:

* Do nothing, so the OP cannot press Ctrl-C on the next check.

* Disable filesystem checks from future boots.

* Switch from Systemd back to SysV Init.

The concept of "going back in time" during an active check, to change a kernel parameter, so it doesn't start to begin with is a strange one. Bugs have been filed, it's even on the TODO for Systemd. So, why not discuss alternatives to the issue until it's fixed?

Re:Skipping fsck during boot with systemd? by foxhill on Sunday December 21, 2014 @09:11AM (#8779774)
i haven't missed anything, you're inventing a scenario based on fragments of knowledge and assumptions.

> is there a mysterious set of commands they came up with to skip an fsck

the final sentence of the opening post.

> I am fully aware that the OP wants to cancel the fsck(8) in progress

well then why are you offering a priori fixes?

> Instead, to prevent additional filesystem checks, there are tools to make this possible

of course there are, but this is entirely irrelevant to the discussion at hand.

> The concept of "going back in time" during an active check, to change a kernel parameter, so it doesn't start to begin with is a strange one.

no strange, impossible. this is my point.

ok, imagine, someone had been in a car accident, and fractured their skull on the window because they weren't wearing their seatbelt. this discussion is akin to a doctor telling the patient (or whoever) that they could wear a seatbelt, and then this wouldn't have happened. just wear a seatbelt in future and you wont fracture your skull. just change the fsck counter/add a kernel boot parameter and fscks wont happen.