Bug 230663 - LSPP: random problems with the python rpm
LSPP: random problems with the python rpm
Status: CLOSED NOTABUG
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: kernel (Show other bugs)
5.0
s390x Linux
urgent Severity medium
: ---
: ---
Assigned To: Red Hat Kernel Manager
Martin Jenner
:
Depends On:
Blocks: 234654 RHEL5LSPPCertTracker
  Show dependency treegraph
 
Reported: 2007-03-01 18:11 EST by Kylene J Hall
Modified: 2007-11-30 17:07 EST (History)
3 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed: 2007-03-26 16:47:55 EDT
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)
Strace output from recently reported passwd issue (15.72 KB, text/plain)
2007-03-15 15:25 EDT, Kylene J Hall
no flags Details

  None (edit)
Description Kylene J Hall 2007-03-01 18:11:49 EST
Description of problem:

I am randomly having problems with the python rpm (at least that is what I think
it is related to).  I was just installing 2 identical partitions on the same
machine with the LSPP kickstart.  One installed and came up fine.  The other
when I was trying to update to the latest packages on the dwalsh people page I
got Segmentation Fault every time I attempted yum update.

Something happened to both partititons and I just tried reinstalling again one
is going along fine the other I get
sh-3.1# yum update
/usr/bin/python: symbol lookup error: /usr/lib64/libpython2.4.so.1.0: undefined
symbol:

Sometimes reinstalling the python rpm with rpm -Uvh --replacepkgs helps
sometimes I resort to reinstalling the whole system.

Other weird things I have gotten in the past that seem like they could be
related.  I'll get errors when trying to update packages about gethomedircon
undefined.  Also sometimes passwd returns passwd: Authentication token
manipulation error (which the last time I saw (yesterday) seemed like like a
glibc corruption issue but also happens randomly during install sometimes.

I'll continue to update with any info or patterns I find.


Version-Release number of selected component (if applicable):


How reproducible:
This has been happening randomly on several of the recent builds but I had been
stalling opening this trying to gather data.  It seems totally random.

Steps to Reproduce:
1. Not sure.
2.
3.
  
Actual results:


Expected results:


Additional info:
Comment 1 Jeremy Katz 2007-03-01 18:54:28 EST
Are you sure your hardware isn't dying?  Please do a memtest86 run to check
Comment 2 Kylene J Hall 2007-03-01 19:13:31 EST
Can you tell me how to do this or point me to some instructions (specifically on
a s390x partion)
Comment 3 Kylene J Hall 2007-03-05 17:27:00 EST
Consensus on the LSPP call earlier today was this is not likely a hardware
problem as I have had it with 2 partitions on one machine and more minorly with
3 partitions on a differnt machine.
Comment 5 Kylene J Hall 2007-03-12 16:12:05 EDT
Ok, now on the "other" machine running a GA VM level is having the issue too. 
First I was getting an Illegal Instruction at line 102 in /etc/init.d/sshd error
when attempting to start sshd on a freshly installed machine.

Now after a second reinstall I get this: /usr/bin/python: symbol lookup error:
/usr/lib64/libpython2.4.so.1.0: undefined symbol:
libsemanage.semanage_install_sandbox: genhomedircon returned error code 127.
Comment 6 Kylene J Hall 2007-03-12 16:36:47 EDT
rpm -Uvh --replacepkg python-2 allowed me to use the setsebool package.
Comment 7 Kylene J Hall 2007-03-13 20:20:25 EDT
Not sure how this is related but when installing today I saw an error when I
updated the kernel during install with yum so I grabbed the package and tried to
update it manually like this.  I also get Illegal Instruction when I attempt to
run zipl.  Attempted to reinstall python (which succeeded) but this problem
persisted.

sh-3.1# rpm -Uvh --replacepkgs kernel-2.6.18-8.1.1.el5.lspp.68.s390x.rpm 
Preparing...                ########################################### [100%]
   1:kernel                 ########################################### [100%]
/sbin/new-kernel-pkg: line 90:  4808 Illegal instruction     $lilo >/dev/null

This is what we get when the kernel attempts to boot:
Booting default (2.6.18-8.1.1.el)...
HCPGIR450W CP entered; disabled wait PSW 000A0000 00000102

And when the other kernel is attempted:
CP VI VMSG 2
Booting linux...
HCPGIR450W CP entered; disabled wait PSW 000A0000 00000102


Presumably b/c zipl is corrupted.

On a reinstall everything went fine.
Comment 8 Kylene J Hall 2007-03-15 15:20:26 EDT
Again not sure how this is related but another data point.  After running
several tests this has begun to happen: 

[root/abat_r/SystemLow@KWUSER1 ~]# passwd -S root
Illegal instruction

Reinstalling python had no affect.

The strace output will be attached.
Comment 9 Kylene J Hall 2007-03-15 15:25:25 EDT
Created attachment 150160 [details]
Strace output from recently reported passwd issue
Comment 10 Kylene J Hall 2007-03-19 17:59:06 EDT
As discussed on the LSPP call I was to try this with the lspp.69 kernel however
it happened on install before I was even able to update any to any of the lspp
packages:

Cut and paste from the kickstart install:

*** Common Criteria configuration kickstart ***


Protection profile (capp or lspp) [capp] ? lspp
Bringing up loopback interface:  [  OK  ]
Bringing up interface eth0:  [  OK  ]
error: /usr/share/rhn/RPM-GPG-KEY: import read failed(-1).

Need to install the certification RPM and updated RPM packages:

lspp-eal4-config-ibm-0.21-1.noarch.rpm


[FIXME] Currently you should also upgrade the following packages
[FIXME] from http://people.redhat.com/sgrubb/files/lspp/
[FIXME]
        acl
        audit audit-libs audit-libs-devel audit-libs-python
        kernel kernel-devel
        libacl libacl-devel
        mcstrans
        openssh openssh-clients openssh-server
        pam pam-devel
        selinux-policy selinux-policy-devel selinux-policy-mls
        selinux-policy-strict selinux-policy-targeted
        vixie-cron

FIXME: updates pending
        rbac-self-test (TBD in config RPM)
        cups (New package pending)

[FIXME] Use a ! escape to put them in the current dir before continuing.

Supply a web URL or a local (absolute) directory name.

If you need to mount a device containing the files,
enter '!' and RETURN to get a shell prompt.

Location [ftp://partners.redhat.com/EAL4_RHEL5/IBM/] ? !

Starting interactive shell, type 'exit' when done
sh-3.1# echo "ARP=no" >> /etc/sysconfig/network-scripts/ifcfg-eth0
sh-3.1# vi /etc/yum.repos.d/sgrubb.repo
sh-3.1# yum update
/usr/bin/python: symbol lookup error: /usr/lib64/libpython2.4.so.1.0: undefined
symbol:
Comment 11 Kylene J Hall 2007-03-19 18:11:53 EDT
Again after re-installing the python rpm I was able to complete my yum update.
However, I went on to have the problem where passwd continuously returns: passwd: 

Authentication token manipulation error
Changing password for user root.
Comment 12 Irina Boverman 2007-03-21 14:50:48 EDT
Per 3/20/2007 meeting with IBM, this bug is urgent, since it blocks IBM's
testing effort. 
Comment 13 Kylene J Hall 2007-03-21 15:16:06 EDT
Latest install on the partition with GA vM level and the kernel wouldn't boot in
enforcing the first time because it couldn't load the policy.  So I booted to
permissive and was attempting to re-install the policy rpms:

[root//@KWUSER3 capp-lspp-rpms]# rpm -Uvh --replacepkgs selinux-policy-*
Preparing...                ########################################### [100%]
   1:selinux-policy         ########################################### [ 25%]
   2:selinux-policy-devel   ########################################### [ 50%]
   3:selinux-policy-mls     ########################################### [ 75%]
/usr/bin/python: symbol lookup error: /usr/lib64/libpython2.4.so.1.0: undefined
symbol:
libsemanage.semanage_install_sandbox: genhomedircon returned error code 127.
   4:selinux-policy-targeted########################################### [100%]
/usr/bin/python: symbol lookup error: /usr/lib64/libpython2.4.so.1.0: undefined
symbol:
libsemanage.semanage_install_sandbox: genhomedircon returned error code 127.

After reinstalling the python rpm the above task was able to complete
successfully. And after a relabel I was able to boot enforcing.
Comment 14 George C. Wilson 2007-03-26 16:29:37 EDT
Can this be closed as not a bug?

Note You need to log in before you can comment on or make changes to this bug.