Please do not reply directly to this email. All additional comments should be made in the comments box of this bug report.
Summary: mkinitrd hack drops 58 cores while creating new initrd
https://bugzilla.redhat.com/show_bug.cgi?id=244328
michal@harddata.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Flag|needinfo?(michal@harddata.co| |m) |
------- Additional Comments From michal@harddata.com 2008-07-16 14:41 EST -------
Can you test this in Fedora 9?
I can but this is not conclusive. The thing is that at the moment I have only x86_64 installations around. A new version of mkinitrd will pick up as LDSO, for use in get_dso_deps(), either files from /lib64/ (that includes desired /lib64/ld-linux-x86-64.so.2 among other things) or will try such "interesting" values for LDSO like /etc/ld.so.conf.d/mysql-x86_64.conf. None of this will attempt to use /lib/ld.so.1.9.5 - which I need for some old binaries and what was really triggering the problem. I am not sure what would happen on i386 installation but on x86_64 /lib/ld* is not attempted at all so the issue is gone.
A workaround for older versions of /sbin/mkinitrd was to exclude in that script "bad" versions of /lib/ld*, which were not needed in the context anyway, thus preventing segfaults.