Announcement

Collapse
No announcement yet.

Systemd 230 Released

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #41
    Originally posted by klapaucius View Post
    Does the Magic SysRq combo work in these cases? It's the quickest way to stop or kill a hanging system, sync the filesystem and shutdown
    Some systems have limited sysrq commands. Most notably some *buntus. In some versions these lulzmakers created kernel with defaults prohibiting anything but Alt-SysRq-B. So you can reboot, but unable to term/kill/sync. EPIC FAIL. Newer versions of *buntus have recognized they're nuts and allowed more commands by default. Erm, I mean, it is tunable in kernel configuration and *buntus screwed it up a bit.

    Comment


    • #42
      Originally posted by phoronix View Post
      Phoronix: Systemd 230 Released

      A new release of systemd is available this weekend...

      http://www.phoronix.com/scan.php?pag...d-230-Released
      After reading all the comments available up to the time I posted this (number 40)....

      Perhaps the problem can be isolated through the following process:
      - install sysv-init (yep, the old original crusty "init" stuff) and remove systemd stuff (as much as possible)
      - make changes as necessary to whatever startup method is used in your Linux OS
      - use the system like you were using
      - drop to a console prompt ans issue "shutdown -h now"
      - see what happens that is different from "systemd" behavior

      Likely problems to occur:
      - you use a GUI and the upstream developers of that GUI (and related stuff) have "tightly coupled" their stuff to "systemd" stuff so the GUI won't start on an "old init-based" system
      - other unexpected breakage due to "tight coupling" to "systemd" and it's stuff

      Proceed at your own risk. If you break your system, don't complain to me.

      Just sayin.....

      Comment


      • #43
        Originally posted by sarfarazahmad View Post
        would it make sense to say that the 90s timeout is a little too much ?
        Just be glad you are not trying to shutdown with a mounted NFS drive that has lost its server. Linux has always been absolutely pathetic at dealing with that, even kill -9 or umount --force can't stop processes accessing a disconnected NFS mount. The only thing that worked was rebooting.

        What systemd now does is count up to 90s, then add 90s more the count, then when that is reached add another 90s to the count-down, and when that is reached add another 90s to the countdown, and ............

        Not sure who is the most incompetent and idiotic in this case though. I think I still put my money on the kernel develops of NFS who apparently are unable to deal with reality, and refuse to accept connections to remote systems can be lost.

        Comment


        • #44
          dude nfs client side has flags which you can use to avoid this (hint: intr, soft) . And this thing isnt specific to Linux, i faced similar issues with AIX as well

          Comment


          • #45
            Originally posted by NotMine999 View Post
            Perhaps the problem can be isolated through the following process:
            If you want to do the hater, do it properly. This post only shows you don't know shit.

            Anyone not liking systemd for their own esoteric reasons are advised to go use true and serious distros (not Devuan, not "without-systemd") that still support modern non-shit non-systemd systems, like openrc, supported on gentoo and the more addomesticated derivative Sabayon linux.


            Comment


            • #46
              Originally posted by nanonyme View Post
              In Windows non-GUI apps that won't shut down in time are just terminated in 30 seconds FWIW
              i.e. same thing, but different and non-editable timeout. but i was talking about non-gui-capable system state, not just not-gui apps
              Last edited by pal666; 24 May 2016, 04:16 PM.

              Comment


              • #47
                Originally posted by startas View Post
                1) Systemd job is to shutdown/restart computer, not to wait for some broken/still working app to hack the universe.
                who told you that bullshit?

                Comment


                • #48
                  Originally posted by startas View Post
                  I totaly disagree with disagreeing - any process must be killed if it takes more than 10 seconds for it to close normally. Shutdown means shutdown, not do ten hours of works and then shutdown, if you want to do something, then do it, and dont shutdown you pc... Linux is just pathetic with these kind of bugs, while windows 10 sucks balls with its too deep integration with internet, retarded flat ui and is basically concentrating on brainwashing, linux still somehow manages to be worse at every aspect...
                  Hey mister, are you aware that you are advocating for an OS that...
                  1. As far as Windows 8, (don't have 10), traps the user that just wanted his pc to shutdown because it forces maj installations that have to be installed (for unknown reasons)? Also, traps the user at start by forcing maj installation the same way (perfect when you just want to get to work fast \o/)?
                  2. As far as Windows 8, still doesn't now how to proper unmount external drives when goes to suspend (making filesystem appear "broken" when you boot on another system)?
                  3. As others pointed, never shutdowns as long as you didn't tell what to do with hanging program (which is actually the behaviour you seem to hate)?
                  4. Whereas Linux has had since ages ago the actual clean process that another forumer describes a few posts above?

                  All the Windows I used actually give many examples of behaviour that is plain bothersome for the user... ^^ Maybe it changed in Windows 10 too (I hope so). After all, Windows users only had to wait 20 years to get audio/video/network working as part of a fresh system install... XD

                  Comment


                  • #49
                    Originally posted by sarfarazahmad View Post
                    dude nfs client side has flags which you can use to avoid this (hint: intr, soft) . And this thing isnt specific to Linux, i faced similar issues with AIX as well
                    I have it set. It doesn't help much. You need to call 'umount --force' several times and hunt down all applications that might as much as look at the mount and kill them to umount. Rebooting is easier, except that now requires a hard power off.

                    Comment


                    • #50
                      Originally posted by carewolf View Post
                      I have it set. It doesn't help much. You need to call 'umount --force' several times and hunt down all applications that might as much as look at the mount and kill them to umount. Rebooting is easier, except that now requires a hard power off.
                      You tried a so-called lazy unmount?

                      unmount -l /share/path

                      This always worked for me on Debian and CentOS.

                      Comment

                      Working...
                      X