Bug 359621
| Summary: | READONLY=yes errors during bootup | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Warren Togami <wtogami> |
| Component: | initscripts | Assignee: | Bill Nottingham <notting> |
| Status: | CLOSED DUPLICATE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | low | Docs Contact: | |
| Priority: | low | ||
| Version: | 9 | CC: | bernie+fedora, dwmw2, rvokal |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2008-08-04 19:19:01 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 188611 | ||
Per the udev - what's in persistent-net.rules for that interface? Is there a persistent net rules file? For /etc/mtab.... one of the reasons we still _have_ /etc/mtab rather than simply making it a symlink to /proc/mounts was because of the special-case in umount(8) for tearing down a loopback device set up be 'mount -oloop'. There are patches which fix that, by setting an 'auto-destruct' flag on the loopback device so it goes away automatically when it's closed/unmounted. kernel: http://marc.info/?l=linux-kernel&m=119361296818388&w=2 util-linux-ng: http://marc.info/?l=util-linux-ng&m=119362955331691&w=2 Are there other reasons for keeping /etc/mtab around? Ah, persistent-net.rules was created by anaconda, containing the interfaces of the host. I have to delete it from the chroot after anaconda installs. Re: mtab - ISTR there being other options preserved in mtab that aren't in /proc/mounts. gid and mode come to mind. hm. uid and mode actually get given to the kernel. What doesn't, for some arbitrary VFAT fs, is 'uhelper=hal' and 'shortname=lower'. No idea what 'shortname=lower' is and why that doesn't show up in /proc/mounts --but 'uhelper' is something else we'd certainly have to deal with. Oh, the meta-option _netdev (used to distinguish network block device mounts) is also not passed into the kernel. Fedora 9 READONLY=yes
Setting up Logical Volume Management: No volume groups found
[ OK ]
find: /var/lock/lvm: Permission denied
find: /var/run/PolicyKit: Permission denied
find: /var/run/mdadm: Permission denied
cp: cannot access '/etc/lvm/archive': Permission denied
cp: cannot access '/etc/lvm/backup': Permission denied
can't create lock file /etc/mtab/~994: Read-only file system (use -n flag to
override)
Mounting local filesystems: mount: devpts already mounted or /dev/pts busy
can't create lock file /etc/mtab~999: Read-only file system (use -n flag to
override)
[FAILED]
Any suggestions?
Did you remember 'setsebool -P allow_mount_anyfile=on'? The chroot has no selinux-policy and is booted with selinux=0. Does that matter? What does the fstab look like? [root@newcaprica ~]# cat /opt/ltsp/i386/etc/fstab /dev/root / auto defaults,noatime 0 0 devpts /dev/pts devpts gid=5,mode=620 0 0 tmpfs /dev/shm tmpfs defaults 0 0 proc /proc proc defaults 0 0 sysfs /sys sysfs defaults 0 0 /dev/root? Why? / is irrelevant anyway since it is mounted by this point. Got rid of the permission denied errors that were the result of NFS root_squash. So there remains only: can't create lock file /etc/mtab/~994: Read-only file system (use -n flag to override) Mounting local filesystems: mount: devpts already mounted or /dev/pts busy can't create lock file /etc/mtab~999: Read-only file system (use -n flag to override) Changing version to '9' as part of upcoming Fedora 9 GA. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping The remaining ugly error message is really a duplicate of 214891 so closing. *** This bug has been marked as a duplicate of bug 214891 *** |
K12LTSP i386 chroot (almost F8 GA packages) booted with selinux=0 selinux-policy is not installed in the chroot. Thin client nfs root boots this chroot, read-only filesystem. /etc/sysconfig/readonly-root: READONLY=yes The below error messages are ugly, but they do not prevent the thin client from working. Setting clock (utc): Tue Oct 30 21:13:23 PDT 2007 [ OK ] Starting udev: udevd-event[459]: rename_netif: error changing netif name eth0 to eth1: Device or resource busy [ OK ] Loading default keymap (us): [ OK ] Setting hostname localhost.localdomain: [ OK ] Setting up Logical Volume Management: No volume groups found [ OK ] find: /var/lock/lvm: Permission denied find: /var/run/mdadm: Permission denied cp: cannot access `/etc/lvm/archive': Permission denied cp: cannot access `/etc/lvm/backup': Permission denied mount: can't find / in /etc/fstab or /etc/mtab Mounting local filesystems: can't create lock file /etc/mtab~775: Read-only file system (use -n flag to override) [FAILED] bash-3.2# ls -ldZ /var/lock/lvm drwx------ root root system_u:object_r:lvm_lock_t:s0 /var/lock/lvm bash-3.2# ls -ldZ /var/run/mdadm drwx------ root root system_u:object_r:mdadm_var_run_t:s0 /var/run/mdadm The permissions don't explain why we're seeing permission denied errors... Starting udev: udevd-event[459]: rename_netif: error changing netif name eth0 to eth1: Device or resource busy [ OK ] udev is behaving badly here. The interface obtained a DHCP lease prior to boot, and copied its DHCP lease with: cp /var/lib/dhclient/dhclient.leases /sysroot/dev/.dhclient-eth0.leases udev shouldn't be trying to rename the interface.