search for: overmount

I just had to deal with a situation where a glusterfs filesystem would not mount because someone had managed to create some files and directories under the mount point. Perhaps this behavior is intentional, but I would prefer that it work like NFS and just overmount. When bringing a cluster down and back up, it is quite possible for overeager users to log in and start working before everything is up. In addition the error message was unhelpful: -bash-3.2# glusterfs --server fs01 /data --volume-name=data glusterfs: could not open specfile -- Matt

..., cst; + struct stat rst, cst, ist; struct statfs sfs; int confd; @@ -186,13 +186,15 @@ const char *run_init(const char *realroo /* Okay, I think we should be safe... */ - /* Delete rootfs contents */ - if (nuke_dir("/")) - return "nuking initramfs contents"; - - /* Overmount the root */ - if (mount(".", "/", NULL, MS_MOVE, NULL)) - return "overmounting root"; + if (!dry_run) { + /* Delete rootfs contents */ + if (nuke_dir("/")) + return "nuking initramfs contents"; + + /* Overmount the root */ + if (mount("....

...e, /* Okay, I think we should be safe... */ - /* Delete rootfs contents */ - if (nuke_dir("/")) - return "nuking initramfs contents"; + if (!dry_run) { + /* Delete rootfs contents */ + if (nuke_dir("/")) + return "nuking initramfs contents"; - /* Overmount the root */ - if (mount(".", "/", NULL, MS_MOVE, NULL)) - return "overmounting root"; + /* Overmount the root */ + if (mount(".", "/", NULL, MS_MOVE, NULL)) + return "overmounting root"; + } /* chroot, chdir */ if (chroot("...

The following patches come from Debian and/or Ubuntu packages of klibc. Ben. Ben Hutchings (1): [klibc] run-init: Add dry-run mode Jay Vosburgh (1): [klibc] ipconfig: Use separate sockets for DHCP from multiple interfaces Mathieu Trudel-Lapierre (1): [klibc] ipconfig: Set broadcast when sending DHCPREQUEST and DHCPDISCOVER YunQiang Su (1): [klibc] mips: setjmp.S: don't

I don't know if it's a bug or a feature, but the filesystem-2.4.0-2.el5.centos rpm won't upgrade cleanly if /home is an NFS filesystem. I sorta-kinda remember this when going from 5.1 to 5.2, but that memory is hazy. -- Paul Heinlein <> heinlein at madboa.com <> http://www.madboa.com/

...-s /ssadrive/Samba /opt/Samba > I would not do that as the tarball has absolute pathnames and it'll only replace the symlink. What you could do is create a filesytem "/opt/Samba" on another VG and mount that. Make sure there is a "/opt/Samba" dir in the /opt filesystem. Overmounting is a clever way to overcome your particularly common scenario of rootvg being too darned small (or overused). I'm copying the list so others can benefit from your scenario. > and then place everything over there? I have planty of space on the other > drives, it is just that the syst...

Hi, I hope this is the right list for initramfs questions. First I noticed: with initrd I can use real-root-device and pivot_root mechanisms. with initramfs only pivot_root works. My init (or linuxrc) scripts end like this: mount -t xfs -n -o ro /dev/mapper/root /new-root umount -n /sys || true umount -n /dev || true umount -n /proc || true cd /new-root pivot_root . initrd exec chroot .

From: Steve Langasek <vorlon@debian.org> The next problem in line was that when trying to clear the initramfs to free memory before overmounting, run-init was getting EPERM instead of EISDIR when calling unlink() on directories. After stepping through both glibc and klibc linked binaries, I've concluded that the issue is simply that klibc is trying to read the errno from the wrong register. Initially, I was concerned this meant unl...

...if (!dry_run) { +- /* Delete rootfs contents */ +- if (nuke_dir("/")) +- return "nuking initramfs contents"; ++ if (!persist_initramfs) { ++ /* Delete rootfs contents */ ++ if (nuke_dir("/")) ++ return "nuking initramfs contents"; ++ } + + /* Overmount the root */ + if (mount(".", "/", NULL, MS_MOVE, NULL)) diff --git a/debian/patches/series b/debian/patches/series index f13f4631..5be39790 100644 --- a/debian/patches/series +++ b/debian/patches/series @@ -1,2 +1,3 @@ resume-backward-compatibility-for-resume_offset.patch kl...

On a train ride to Bruxelles, brought out my axe and directly attacked run_init(8). run_init(8) is dead, long live switch_root(8). The next run on switch_root(8) involves fdopendir, so another push for the upcoming stdio 1.6 branch. The following is boot tested with initramfs-tools, kinit(8) tests would very much be appreciated!? Michal Suchanek (1): [klibc] switch_root: Fix single file

...V_INITRD if you want to pass an initramfs archive via the bootloader. During boot the kernel unpacks the built in archive first, then the one passed by the bootloader (if either of them exist of course). Initramfs archives are unpacked directly into the rootfs, a ramfs which is usually immediately overmounted by the user-visible root. If an initramfs is found the kernel does not mount a root filesystem and starts /init in the initramfs. /init is responsible for mounting the root and exec'ing the normal init. Once /init has performed all necessary set-up tasks and initramfs is no longer needed, th...

...... */ if (!dry_run) { - /* Delete rootfs contents */ - if (nuke_dir("/")) - return "nuking initramfs contents"; + if (!persist_initramfs) { + /* Delete rootfs contents */ + if (nuke_dir("/")) + return "nuking initramfs contents"; + } /* Overmount the root */ if (mount(".", "/", NULL, MS_MOVE, NULL)) -- 2.21.0.392.gf8f6787159e-goog

...... */ if (!dry_run) { - /* Delete rootfs contents */ - if (nuke_dir("/")) - return "nuking initramfs contents"; + if (!persist_initramfs) { + /* Delete rootfs contents */ + if (nuke_dir("/")) + return "nuking initramfs contents"; + } /* Overmount the root */ if (mount(".", "/", NULL, MS_MOVE, NULL))

..., newroot); if (mount("none", newroot, "tmpfs", 0, getmemsize()) < 0) serror("mounting tmpfs"); if (chdir(newroot)) serror("entering new root"); move_dir("/"); if (mount(".", "/", NULL, MS_MOVE, NULL)) serror("overmounting root"); if (chroot(".") || chdir("/")) serror("chroot"); /* remove copy of this program */ unlink("/init"); exec_init(); return 1; } -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: applic...