Re: ES3 update 8 has severe problems with "pvm"

Taroon Users:

On Fri, 21 Jul 2006, James Leinweber wrote:

> On any server with the "pvm" package installed, which seems to creep
> in by default, running "up2date -u" on 2006-July-20 dies with the
> message:
>
>  error: unpacking of archive failed on file
>    /usr/share/pvm3/bin/LINUXI386: cpio:  rename failed - Is a directory
>
> This leaves the system in a very unfortunate state, with the updates
> partially applied, and about 50 RPM's missing in action, including
> rather important things such as "openssh-server".  With a lot of
> hand work using "rpm", 'find ... -name "*.rpm*" -print' and the like I
> was able to get a couple of systems past that.
>
> a) is anyone else having this problem?

July 20th I was Upgrading a Remote WS3-U7 System (with all previous 
updates already applied), using the Graphical version of UP2DATE via 
NoMachine's NX Client.  Part way through the Upgrade there was an error 
related to the pvm update (in the Graphical Mode I did NOT see/remember 
the detailed Error Message which James saw/noted), so I restarted UP2DATE 
to complete the WS3-U8 Upgrade (which generated no further errors and 
completed with RHN Icon going back to its "comforting" Blue Checkmark)...

After closing the NoMachine NX Session I attempted to Log back in using 
the Secure Shell Client (to reboot following the Kernel Updates included 
within the WS3-U8 Upgrade), but found no response on the SSH Port (or from 
the NX Server, which piggy-backs on the SSH Port)!!!  After getting the 
System's Owner to temporarily enable the Telnet Port at the Firewall 
Level, I could see all Links within the Sys-V Run Level Tree for the SSH 
Server had been removed???  Prior to that the Owner had tried rebooting 
the System: both before and after that reboot the SSH Port was NOT 
responding, so I assume the Upgrade had killed (and failed to restart) the 
SSH Server following the SSH Update???

> b) BEWARE!  You probably want to avoid running "up2date" on systems
> with "pvm" installed unless you are a masochist.

According to the rpmpkgs Log Files pvm-3.4.4-22.i386.rpm was installed on 
the system at the WS3-U7 Level and the WS3-U8 Upgrade had attempted to 
update using pvm-3.4.5-6_EL3.i386.rpm (which is NO longer in the system 
and I did NOT explicitly remove it myself)???  There were only a handful 
of *.rpmnew Files on the System and they all seems older than the WS3-U8 
Upgrade anyway...  The effected system is a RHEL WS3 Pre-installed DELL 
System, so possibly DELL had included the pvm RPM (PVM is NOT something I 
Installed, Require or Use)???

> c) applying update 8 using Redhat Network went very smoothly for me on
> a system which didn't have "pvm" installed.
>
>I note that
>   rpm -q --whatrequires pvm
> didn't suggest any dependencies, so those of us not doing parallel
> distributed programming may want to do "rpm -e pvm; up2date -p" as
> a workaround before trying to apply update 8.

Locally I tried a WS3-U7 to WS3-U8 Upgrade within a VMWare Virtual 
Machine: the Upgrade proceeded smoothly and the SSH/NoMachine NX Servers 
were NOT effected...  NOTE: The pvm RPM does NOT appear within the VMWare 
Guest's up2date Log File, so I assume it was NOT there before the WS3-U7 
to WS3-U8 Upgrade???

Whether UP2DATE halting on that ill fated PVM Update was the cause of the 
SSH Server Lose, I can NOT be sure, but having lost SSH/NoMachine NX 
Access to a System 100 Miles away was NOT my idea of "Fun"...  Hopefully 
this was just a "Bug" which will be fixed and NOT a "new" "Feature", or 
NOT one I would expect at the Enterprise Level???

Lawrence Houston  --  ([email protected])

--
Taroon-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/taroon-list