Subject: Re: RH9 / glibc update (fwd)
From: Rob Cermak (cermak@sfos.uaf.edu)
Date: Fri Apr 11 2003 - 09:43:31 AKDT
Hi,
Pardon the slow response, I'm in digest mode :)
In the history of all the glibc updates I've done on RH systems, I've
never had this big a failure. I've never had to do a telinit 1 or
reboot. I realize programs have to be restarted to use the new glibc
because of old copies being kept in memory until restarting...
After I installed the update, all core programs failed to operate and
'segfaulted' down to reboot. I was hoping that 'rpm' would still
function: No! Segfault. Can't even back down to the old glibc :(
I did the hard reboot and poof, the kernel boots and it starts init;
nothing. The base init processes are said to be respawning too rapidly
and the whole machine is hosed. I guess they were spawning and
segfaulting...
Now... it could be this particular motherboard.
P4 ASUS P4PE Socket 478, Intel 845PE
RH 8 and RH 9 operated just fine before on this box.
I have another test system at home Dell that I'll perform the same
experiment.
--------------------------------------------------------------------------
From: Jim Gribbin <jgribbin@alaska.net>
To: Rob Cermak <cermak@sfos.uaf.edu>
Cc: AKLUG <aklug@aklug.org>
Subject: Re: RH9 / glibc update
I don't know if it's the same on RH 9, but there is also a glibc update
errata for RH 8 that states that after the update the system must be
rebooted or all programs restarted.
Jim Gribbin
---------- Forwarded message ----------
Date: Fri, 11 Apr 2003 08:54:52 -0800 (AKDT)
From: volz <volz@koyukuk.at.uaa.alaska.edu>
To: aklug@aklug.org, cermak@sfos.uaf.edu
Subject: Re: RH9 / glibc update
Rob:
Have you gotten any resolution on your glibc update? Did you do the telinit 1
to restart all programs? Redhat has this in their updated errata:
https://rhn.redhat.com/network/errata/errata_details.pxt?eid=1543
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
[Updated 9 April 2003]
Some bugs newly introduced by the Red Hat Linux 8.0 errata packages
has been discovered, such as wine problems or problems with debugging
threaded applications.
Once the glibc upgrade has been completed, you must either reboot the
system or restart all programs on the system (for example, by using telinit
1 and switching back to the old runlevel). Rebooting the system or
restarting the system programs is necessary to avoid vulnerable glibc
copies in memory and because changes in NSS internal interfaces which mean
that one cannot mix old NSS modules or libresolv in an application with
upgraded NSS modules or libresolv in one running application.
If sshd is running so that the other services can be restarted remotely or
for a remote reboot during an unattended glibc upgrade, glibc will also
restart sshd.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
I realize it doesn't specifically mention RH9.
The reason I ask is I don't want to hose my system if Redhat hasn't got this
patch worked out.
- Karl Volz
Research Associate
Alaska Experimental Forecast Facility
2811 Merrill Field Dr.
Anchorage AK 99501
(907) 264-7427
volz@aeff.uaa.alaska.edu
>
>Anyone else having trouble with RH9 and the glibc update. Once I updated
>with the new glibc, the system is now completely hosed. Most commands
>fail with segmentation violation.
>
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
This archive was generated by hypermail 2a23 : Fri Apr 11 2003 - 09:43:33 AKDT