Something is causing abrt-dump-oops to start up in live media creation chroots. This causes the mount to stay busy, which in turn causes livecd-creator to lazy unmount it, which leads to a corrupt image. ;( I'm not sure what starts this, but it should not start in chroots. ;( Livecd-creator outputs: Unmounting directory /var/tmp/imgcreate-L7sNK5/install_root umount: /var/tmp/imgcreate-L7sNK5/install_root: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) # fuser -m /var/tmp/imgcreate-L7sNK5/install_root /var/tmp/imgcreate-L7sNK5/install_root: 13892rce # ps -axuww | grep 13892 root 13892 0.0 0.0 3488 612 ? Ss 22:11 0:00 abrt-dump-oops -d /var/spool/abrt -rwx /var/log/messages # ls -l /proc/13892/cwd lrwxrwxrwx. 1 root root 0 May 12 22:22 cwd -> /var/tmp/imgcreate-L7sNK5/install_root # pstree | grep dump-oops systemd-+-abrt-dump-oops Note that removing this package on the host you are composing doesn't help. it happens if the package is installed on the live media in the chroot you are composing. Happy to provide more info or debugging.
-- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
Scriptlets from this subpackage are likely the issue: postinstall scriptlet (using /bin/sh): if [ $1 -eq 1 ]; then /sbin/chkconfig --add abrt-oops fi preuninstall scriptlet (using /bin/sh): if [ "$1" -eq "0" ] ; then service abrt-oops stop >/dev/null 2>&1 /sbin/chkconfig --del abrt-oops fi #systemd (not tested): if [ "$1" -eq "0" ] ; then /bin/systemctl stop abrt-oops.service >/dev/null 2>&1 || : /bin/systemctl disable abrt-oops.service >/dev/null 2>&1 || : fi posttrans scriptlet (using /bin/sh): if [ "$1" -eq "0" ]; then # this is a tmp hack to set-up the ccpp hook when updating # from 1.x to 2.x without restarting service abrt-oops restart >/dev/null 2>&1 || : fi #systemd if [ "$1" -eq "0" ]; then /bin/systemctl try-restart abrt-oops.service >/dev/null 2>&1 || : fi
The posttrans script is wrong.
(In reply to comment #3) > The posttrans script is wrong. From notting ... "fairly sure that %posttrans should be %postun, and the refcounting is wrong"
Created attachment 498796 [details] updated scriptlets Fixes here: 1) if [ "$1" -eq "0" ] is a no-op for %posttrans - the first argument is always 0. (http://fedoraproject.org/wiki/Packaging:ScriptletSnippets#Syntax) 2) We need to call condrestart, not restart - restart *will start* the service, which we don't want unilaterally. 3) Eliminate the systemctl calls; service <foo> condrestart will automatically redirect when necessary. As written, the scriptlets restart the service twice.
Basically any live-based release criteria we have would be blocked by the lack of live media. Clearly +1 Blocker from me.
+1 from me (actually I proposed it). other votes? kklic is working on an update for this. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
abrt-2.0.2-5.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/abrt-2.0.2-5.fc15
Package abrt-2.0.2-5.fc15: * should fix your issue, * was pushed to the Fedora 15 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing abrt-2.0.2-5.fc15' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/abrt-2.0.2-5.fc15 then log in and leave karma (feedback).
abrt-2.0.2-5.fc15 has been pushed to the Fedora 15 stable repository. If problems still persist, please make note of it in this bug report.
*** Bug 681293 has been marked as a duplicate of this bug. ***