Bug 41202 - Upgrade to 4.6.1-RELEASE-p3 breaks remote ssh login until the next reboot
Summary: Upgrade to 4.6.1-RELEASE-p3 breaks remote ssh login until the next reboot
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: misc (show other bugs)
Version: 4.6.1-RELEASE
Hardware: Any Any
: Normal Affects Only Me
Assignee: freebsd-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-07-31 14:20 UTC by Diomidis Spinellis
Modified: 2003-07-13 01:30 UTC (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Diomidis Spinellis 2002-07-31 14:20:01 UTC
After the make installworld of a  4.6.1-RELEASE-p3 upgrade the machine will stop accepting incoming ssh connections due to loadable module mismatches:
Jul 31 15:34:58 istlab sshd[11976]: unable to dlopen(/usr/lib/pam_skey.so)
Jul 31 15:34:58 istlab sshd[11976]: [dlerror: /usr/lib/pam_skey.so: Undefined symbol "pam_set_option"]

Fix: 

A reboot fixes this problem. It might be worthwhile to notify 
potential upgraders.  
Those performing a remote upgrade will want to keep their ssh terminal 
session open and execute a reboot after the installworld, 
otherwise they will have to get local access to the machine's console.
Comment 1 corecode@corecode.ath.cx 2002-07-31 15:58:45 UTC
On Wed, 31 Jul 2002 06:18:16 -0700 (PDT) Diomidis Spinellis wrote:

> >Fix:
> A reboot fixes this problem. It might be worthwhile to notify 
> potential upgraders.  
> Those performing a remote upgrade will want to keep their ssh terminal
> session open and execute a reboot after the installworld, otherwise
> they will have to get local access to the machine's console.

as the handbook states in 19.4:

19.4.8 Reboot into Single User Mode
[...]
19.4.9 Install the New System Binaries  [i.e. installworld]

you should do an installworld only while running the system in single
user mode. this way no collisions happen with network, pam, .so's or
whatever.


-- 
/"\   http://corecode.ath.cx/#donate
\ /
 \     ASCII Ribbon Campaign
/ \  Against HTML Mail and News
Comment 2 Peter Pentchev 2002-07-31 16:38:10 UTC
On Wed, Jul 31, 2002 at 08:00:17AM -0700, Simon 'corecode' Schubert wrote:
>  On Wed, 31 Jul 2002 06:18:16 -0700 (PDT) Diomidis Spinellis wrote:
>  
>  > >Fix:
>  > A reboot fixes this problem. It might be worthwhile to notify 
>  > potential upgraders.  
>  > Those performing a remote upgrade will want to keep their ssh terminal
>  > session open and execute a reboot after the installworld, otherwise
>  > they will have to get local access to the machine's console.
>  
>  as the handbook states in 19.4:
>  
>  19.4.8 Reboot into Single User Mode
>  [...]
>  19.4.9 Install the New System Binaries  [i.e. installworld]
>  
>  you should do an installworld only while running the system in single
>  user mode. this way no collisions happen with network, pam, .so's or
>  whatever.

This advice may often be disregarded, *BUT* not lightly!  One must
always be aware of the changes made to the system between rebuilds, if
one chooses not to install in single user mode, or not to reboot after
the installation is complete.  The only ways I can think of to "always be
aware" is to either track the changes via cvsweb or something similar,
or read the src/UPDATING and similar files, *or* run the 'periodic
security' command after the installation and take a good look at the
differences it encounters.  Special care should be taken with programs
that change, and especially long-running daemons that have changed, such
as sshd, or libraries that long-running daemons use.  If such a daemon
or a library should change during an installworld, it is almost certain
that you have to restart the daemon to make sure that you are really
running the latest version.

I wonder if some of the information in what I just said should not go
into the handbook..  I just might write something up and post it to -doc
for review soonish :)

If you are wondering what should be a legitimate reason not to go down
into single-user mode, here's a hint - colocation :)

G'luck,
Peter

-- 
Peter Pentchev	roam@ringlet.net	roam@FreeBSD.org
PGP key:	http://people.FreeBSD.org/~roam/roam.key.asc
Key fingerprint	FDBA FD79 C26F 3C51 C95E  DF9E ED18 B68D 1619 4553
If wishes were fishes, the antecedent of this conditional would be true.
Comment 3 Kris Kennaway freebsd_committer freebsd_triage 2003-07-13 01:30:15 UTC
State Changed
From-To: open->closed

This is expected behaviour