DO NOT UPGRADE YOUR HARDY BOXES
March 12th, 2008 | Published in Personal | 33 Comments
I just learned the hard way. I updated all 3 about 15 minutes ago, and lets just say, none of them work. I can get to TTY, but can’t log in. malloc this and malloc that everywhere.
Thank God for my Foresight install! Otherwise, like I just told Ken Vandine, if it weren’t for Foresight, I would be watching TV right now :p So I guess while I have Foresight up and running, KDE 4.0.2 as well baby, I might as well work on some Foresight bugs now.
Talk about irritating, I was working on my 5-a-day and my really-fix-its. Now I will work on a small https issue, some rebranding for Foresight KDE 4, and maybe even some documentation. Get ready Ken, here I come! Prepare to upload 
/me ssh’s into Og’s box and does ‘sudo conary update group-kde-dist’
March 12th, 2008 at 23:29:57 (#)
too late for me
yes i just updated my hardy 5 minutes ago and it is now dead. i am typing this from a live cd boot and reinstalling gutsy.
March 12th, 2008 at 23:33:32 (#)
ramil, gah, that sucks….You and I aren’t the only ones. There is some twitter traffic complaining about the same thing. libc6 issue obviously. I am hoping that I will be able to fix it when libc6 gets fixed. It will have to be from a live cd, because I cannot log in at all. If it comes down to me having to format and reinstall, I will be super pissed. I have had Kubuntu on my laptop for 2 years now w/o having to reformat or reinstall, I have been dist-upgrading since Dapper and I don’t want to lose that. If I do, then I don’t know what route I will go with a new install unfortunately.
March 13th, 2008 at 00:52:25 (#)
Thanks a lot for the warning !
I’ll wait for a few days for this bad bug to be fixed (I hope they’ve pull out the faulty updates already though).
March 13th, 2008 at 00:53:31 (#)
I just upgraded my box when I saw this.
There goes the work on my commute…
March 13th, 2008 at 00:55:13 (#)
I didn’t have a problem with all 296 updates earlier today. All went well reboot was required and rebooted fine for me.
March 13th, 2008 at 01:13:58 (#)
Umm… got any bugs to track? I’ve updated hardy a few times today. What broke?
March 13th, 2008 at 01:14:21 (#)
No problems here; I just upgrade Hardy and everything works.
March 13th, 2008 at 01:21:37 (#)
I guess it must be due to https://bugs.launchpad.net/ubuntu/+source/glibc/+bug/201673
Fortunately, it’s been repealed (the binary packages are 403ing).
March 13th, 2008 at 01:46:37 (#)
I added some comments to the related bug, I hope it will help.
March 13th, 2008 at 03:51:35 (#)
I have always wondered if anybody tests these things before they reach the repos. Surely it would take just 1 minute to spot this terrible bug and save us from the disaster. Yet they seem to refuse to add minor software updates (updated pidgin for example) for ‘stability sake’.
If you ask me, bugs like this one is not acceptable, even on alpha stage. Not in linux-world. More so when preventing it only requires 30 seconds and a pair of working eyeballs.
March 13th, 2008 at 04:19:28 (#)
is it safe to upgrade now?
March 13th, 2008 at 04:59:09 (#)
@ak,
Alpha by definition is unstable right? People have faced repeated crashes, unbootable systems while testing alpha releases. How is this different?
And if you read other comments, not everyone has faced the problem. Which is strong suggestion that uploader might not have faced the problem when he tested the package. Just because you wonder ‘if anyone tests it’ does not mean that ‘no one test packages before upload’.
By the way why are you confusing ‘refusal to upload new verson for stability sake’ with this one?
March 13th, 2008 at 05:16:15 (#)
You know, this bug and the kernel update a few days ago that broke sound for everyone makes me wonder how packages are tested…sure, it’s an alpha but still. If bugs affects so many system that it effectively is universal in scope one has to wonder if someone bothered to install it on even one test-system and rebooting before it was pushed into the repos.
Honestly, it almost sounds as if it was assumed it would be working.
March 13th, 2008 at 07:01:04 (#)
She’s steamin’ along fine for me.
March 13th, 2008 at 07:17:53 (#)
Onkar, it seems that Jonas understands what I’m talking about.
When a bug has a hit rate of 80% (maybe more) it really makes me wonder if it was even tested once. I know what Alpha means, but I have come to expect more of Linux, even in development stages. I didn’t ask for everything to be perfect. I’m not crying foul every time some dependencies get screwed. But this is not your average dependency-bug. This one completely and utterly kills just about everything. So, my question I think is valid. A bug like this reaching the repos makes you wonder how much it was tested, if at all. Doom’s-day warning signs are no substitute for lack of even-the-slightest preparation/testing. I feel that with a tiny bit more effort put in testing, everything would be okay now. “Alpha” doesn’t mean “no quality control.” Maybe in Windows it does. Not in Linux.
This in contrast with minor application upgrades which are refused. To my experience, these little updates rarely bring in unsatisfiable dependencies, and never ever render a system useless.
Perhaps the “does it annihilate everything?” factor should play a larger part of package acceptance/refusal decisions?
March 13th, 2008 at 07:38:18 (#)
hehehe Watch out for the flying crutches!
March 13th, 2008 at 07:43:42 (#)
I’m locked out also..without any ttys, I feel a little helpless.
Is there any alternative to formatting?
March 13th, 2008 at 09:20:54 (#)
Good solution is here.
https://bugs.launchpad.net/bugs/201673
March 13th, 2008 at 09:23:13 (#)
I ran updates last night and then shutdown and went to bed. I wonder what’s waiting for me when I get home
. Oh joy. On the bright side it looks like there is a fix and I have a second hard drive with gutsy any way. Thanks for the heads up.
March 13th, 2008 at 09:35:03 (#)
Thanks Marcin!
March 13th, 2008 at 10:18:51 (#)
@Marcin: thanks for the link there. I found the solution last night as well in the forums before going to sleep but was to tired to link it right away.
As for the testing part, those who are running Hardy are testing it. Everyone is human and we tend to make a mistake every now and then. Now if this were a stable release, well that would be a different story of course. Stuff like this is expected during a development cycle.
March 13th, 2008 at 10:49:45 (#)
The best solutiuon is to rollback to a previous version of libc6 package.
Do this
1) Boot up with a LiveCD. You can use any available livecd; Gutsy or Hardy livecd or Suse livecd etc.
2) Mount the Hardy’s damaged filesystem to /media/. Here is an example what I did.
$ sudo -s
# mkdir /media/sdb1
# mount /dev/sdb1 /media/sdb1
3) Perform a chroot on that mounted partition. All command in chroot will affect on the chroot jailed box.
# chroot /media/sdb1
4) Download an older version of libc6 page. Do
# wget http://us.archive.ubuntu.com/ubuntu/pool/main/g/glibc/libc6_2.7-5ubuntu2_i386.deb
5) Install the package
# dpkg -i libc6_2.7-5ubuntu2_i386.deb
It will downgrade the libc6 from version 2.7-9ubuntu1 to 2.7-5ubuntu2.
6) exit
7) Reboot and everything is ok.
March 13th, 2008 at 11:27:46 (#)
i think it’s save to upgrade from main server repo now with new release of 2.7-9ubuntu2
IMHO, i dont think we should blame anyone for this bug, if it’s aint broken,then it’s not development branch. we should’t be scared to this kind of problem while using development version.
anyhow, thanks to this bug now i know if anything happen to my ubuntu, i still can fix it from live cd..with many ways lol
reinstall is shortest path for who dont have patient and time to think
March 13th, 2008 at 12:00:56 (#)
Glad I read this before updating last night. Thanks for the warning
March 13th, 2008 at 12:12:49 (#)
@Joe Smith: exactly why I posted. Tried to provide some warning. I posted it late here in Chicago, so I knew the people on the other side of the planet would have got the warning. I think most of us in the US got the warning to late though
March 13th, 2008 at 14:26:52 (#)
I did yesterdays update and it was fine, but today there are 20 more updates now I’m reluctant to do those updates. Will someone tell me when and how many the updates were that broke the system? As I posted last night from a windows box at work “All 296 updates yesterday afternoon went fine” there are 20 more today. Should I update those?
March 13th, 2008 at 14:29:25 (#)
Yesterdays update worked, but today there are 20 more updates now I’m reluctant to do those updates. Will someone tell me when and how many the updates were that broke the system? Should I update those?
March 13th, 2008 at 15:32:28 (#)
I just did another 36 updates after the 20 and rebooted. Everything is fine still. I think the problem has been SOLVED. My blog http://nudepenguin.net has the screenshots.
March 13th, 2008 at 15:39:11 (#)
This is such an awesome example of how UBUNTU trumps microsuck windows. As I have posted to my blog if this were a windows update that was breaking systems it would have took at least 2 months for microsuck to release an update to the update that broke the system. It took UBUNTU less than 24hrs.
March 14th, 2008 at 11:13:36 (#)
Can I ask, is the library fixed or has it just been pulled? Because I upgraded 2 systems over the course of the last 2-days and they both have libc6 v.2.7-9 installed. I haven’t rebooted, and wondering if I should downgrade to version v.2.7-5 before I reboot. Thx
March 14th, 2008 at 11:21:03 (#)
I guess its fixed as the launchpad has this…. “Fixed packages (version 2.7-9ubuntu2) are now on archive.ubuntu.com.”
I will cross my fingers on the reboot.
March 15th, 2008 at 08:07:37 (#)
I had quite a handful of installs as well.. All would work fine for like a couple of hours or so.. And then one update spoils all fun… My Xserver crashed like a million times and the notebook LCD used to flicker.. Any idea what is the issue ???!!
March 19th, 2008 at 01:47:49 (#)
Same thing happened to me.. Back on 7.10 after a couple of hours of re-installing / re-configuring… Guess I’ll be waiting for the official release now…