Don't install glibc-2.12.90-13

[Hansvon]

I'm not sure if this affects everybody but on my notebook, after having installed this update, I couldn't run any program (instant segmentation fault). I couldn't reboot either (init segfault).
Issue is confirmed by somebody else on bodhi: https://admin.fedoraproject.org/updates/glibc-2.12.90-13

Please be careful with this update. I had to boot from a live CD and install previous glibc manually...

[tox]

it looks to be a regression something to do with grep-2.7 ? http://lists.fedoraproject.org/pipermail/devel/2010-September/143480.html

just on the off hand, downgrade grep to 2.6, but keep that glibc

[is977]

I'm not sure if this affects everybody but on my notebook, after having installed this update, I couldn't run any program (instant segmentation fault). I couldn't reboot either (init segfault).
Issue is confirmed by somebody else on bodhi: https://admin.fedoraproject.org/updates/glibc-2.12.90-13

Please be careful with this update. I had to boot from a live CD and install previous glibc manually...

I have same problem on my work machine with c2duo. On my home machine with Athlon X2 after update system work fine.

[Ales]

Same here, instant segmentation faults everywhere. Had to use Live CD to roll back to glibc and glibc-common-2.12.90-11 (from ...-13).

Intel c2duo T9300, Fedora Beta RC3 x86_64.

Just posting to add +1 to a word of caution for everybody, DO NOT UPGRADE GLIBC* UNTIL THIS IS RESOLVED.

[Boricua]

Hi, unfortunately, it is already too late for me. Could please anyone posts some instructions as to how to roll back the glibc packages using the live dvd? I'm using F14 64 bit.

[Hansvon]

Post deleted.
Please use yum --installroot=... downgrade glibc* (or history undo as explained below by LightDot.)

[Boricua]

I manually extracted the rpm in my root partition (I couldn't do it properly with chroot and rpm/yum command, chroot segfault'ed). Maybe, you should wait for a better method.

Replace sdaX with your root partition and extN with its type (ext3,ext4). Maybe, you don't need the devel and header packages, and you don't need x86_64 packages if you're running i686... You need to be in /mnt/sda before executing rpm2cpio...

su -
mkdir /mnt/sda
mount -t extN /dev/sdaX /mnt/sda
cd /mnt/sda

yumdownloader --disablerepo=updates-testing glibc.i686
yumdownloader --disablerepo=updates-testing glibc.x86_64
yumdownloader --disablerepo=updates-testing glibc-devel.i686
yumdownloader --disablerepo=updates-testing glibc-devel.x86_64
yumdownloader --disablerepo=updates-testing glibc-common
yumdownloader --disablerepo=updates-testing glibc-headers

rpm2cpio glibc-2.12.90-11.i686.rpm | cpio -iuvd
rpm2cpio glibc-2.12.90-11.x86_64.rpm | cpio -iuvd
rpm2cpio glibc-devel-2.12.90-11.i686.rpm | cpio -iuvd
rpm2cpio glibc-devel-2.12.90-11.x86_64.rpm | cpio -iuvd
rpm2cpio glibc-headers-2.12.90-11.x86_64.rpm | cpio -iuvd
rpm2cpio glibc-common-2.12.90-11.x86_64.rpm | cpio -iuvd

One last important thing. A few configuration files belong to glibc and glibc-common, you may want to backup them

glibc /etc/gai.conf
glibc /etc/ld.so.cache
glibc /etc/ld.so.conf
glibc /etc/ld.so.conf.d
glibc /etc/localtime
glibc /etc/nsswitch.conf
glibc /etc/rpc
glibc-common /etc/default
glibc-common /etc/default/nss
glibc-common /etc/gai.conf

Thanks Hansvon, I will give it a try.

[LightDot]

I also couldn't chroot due to segfault and ended up using yum in a different way. This is the way I did it:

- boot Live CD
- mount your root partition on hdd
- and then in the Live CD terminal:

su -
yum --installroot=path_of_your_mounted_root_partition history

This will display all your previous yum transactions. The glibc update should be the last or among the last ones. Remember the ID of the transaction you think it is and check with the next command (substitute ID with your actual ID number):

yum --installroot=path_of_your_mounted_root_partition history info ID

Once you confirm it's the ID you need, you can undo the transaction:

yum --installroot=path_of_your_mounted_root_partition history undo ID

The undo might not be able to roll back all the packages, but it should be able to roll back atleast glibc and glibc-common. If there are errors and you end up with duplicate glibc and glibc-common packages, do the following (first command should report only two installed packages, if it shows that duplicates are installed, remove the newer ones):

yum --installroot=path_of_your_mounted_root_partition info glibc glibc-common
yum --installroot=path_of_your_mounted_root_partition remove glibc-2.12.90-13.x86_64 glibc-common-2.12.90-13.x86_64

Then you'll be able to boot your standard Fedora 14 Beta install and complete other yum transactions that failed.

[Boricua]

Ok, I will try then LightDot's suggestion. Thanks to both.

[lmcogs]

Pity I did not read this today. Same problem here when updating

Luckily I had another partition which I booted into and followed
https://bugzilla.redhat.com/show_bug.cgi?id=638091 and followed Jeffrey C. Ollie advice and I was able to boot again.

then when I did an update glibc I still see its 2.12.90-13. For a potentially disastrous it's slow to be fixed.

[nimnull22]

There is glibc-2.12.90-14 in http://koji.fedoraproject.org/koji/buildinfo?buildID=197187.
You can try it if you want.

[dd_wizard]

This odd. According to bodhi (https://admin.fedoraproject.org/updates/glibc-2.12.90-14), -14 has been pushed to stable. However, after "sudo yum clean all", the only update I see is -13 from updates-testing. Any ideas what's up with that?

dd_wizard

[nimnull22]

I always downloaded from koji by my self. I think update has not uploaded to all servers yet.

[Vector]

If you DO install it (because you didn't read this thread first, ahem, ME), then here's one way to fix it in cases where you can't even boot: http://forums.fedoraforum.org/showthread.php?t=252170

[Hansvon]

Your solution works too (rpm --force) but downgrading with yum is cleaner (yum --installroot=.. downgrade glibc*).

[Vector]

Oh, i agree, i just didn't know how to do that. It's a shame, i've been using Fedora for a long time, but i'm still just a web developer --my system admin skills are far weaker than they should be, because i don't study/practice that stuff. So, i was lucky to get my system bootable again :D. So, it's at least an alternative, for "worst case scenarios".

Thanx

[jonathan47]

I also couldn't chroot due to segfault and ended up using yum in a different way. This is the way I did it:

- boot Live CD
- mount your root partition on hdd
- and then in the Live CD terminal:

su -
yum --installroot=path_of_your_mounted_root_partition history

This will display all your previous yum transactions. The glibc update should be the last or among the last ones. Remember the ID of the transaction you think it is and check with the next command (substitute ID with your actual ID number):


I'm not getting any history after mounting the root partition. I must be mounting the wrong way.

My F14 is on dev/sdb, which has a boot partition at sdb1 and LVM at sdb2. The root file system is at /dev/dm-3.

Using the Live USB of F14, I mounted sdb2 and, when that didn't produce a yum history, I mounted dm-3, still with no luck. Here are the mount commands I used as root:

mkdir /mnt/sdb
mount -t ext4 /dev/sdb2 /mnt/sdb

and
mkdir /mnt/sdb2
mount -t ext4 /dev/dm-3 /mnt/sdb2

Any help is appreciated. Thank you.

[cgrim]

What about version 2.12.90-14? Have someone tried it? Is it ok? Thanks ;)

[lmcogs]

Yum updated to 14 and it works fine

[chepioq]

I confirm, I install glibc-2.12.90-14 and that work.

[cgrim]

Thank you for your info, it really works ;-)

[downer]

I wondered what was going on. I couldn't get to this forum, but using the install cd and "chroot /mnt/sysimage" several times with the up arrow and then enter very quickly worked, and then "yum update" the same way eventually worked.

[Boricua]

I'm not getting any history after mounting the root partition. I must be mounting the wrong way.

My F14 is on dev/sdb, which has a boot partition at sdb1 and LVM at sdb2. The root file system is at /dev/dm-3.

Using the Live USB of F14, I mounted sdb2 and, when that didn't produce a yum history, I mounted dm-3, still with no luck. Here are the mount commands I used as root:

mkdir /mnt/sdb
mount -t ext4 /dev/sdb2 /mnt/sdb

and
mkdir /mnt/sdb2
mount -t ext4 /dev/dm-3 /mnt/sdb2

Any help is appreciated. Thank you.

I got a similar issue and asked a friend for help. He found out the instructions for mounting available at the bugzilla's link available somewhere else in this thread (the current bug report) did the trick. Hope it helps you as well.