Martin Paul | 3 May 2012 11:06
Picon
Picon
Favicon

new "pkgdep" command

Not directly related to PCA, but might be handy: In the README of the new 
revision of the "Install and Patch Utilities Patch" for Solaris 10 (119254-84, 
119255-84) I spotted this:

   5012345 request for tool to list package dependencies

And indeed, the patches include a new script /usr/sbin/pkgdep, which prints 
dependencies of a given package. E.g.:

$ pkgdep -v SUNWgzip
SUNWcar - Core Architecture, (Root)
SUNWcsd - Core Solaris Devices
SUNWcsl - Core Solaris Libraries
SUNWcsr - Core Solaris, (Root)
SUNWcsu - Core Solaris, (Usr)
SUNWkvm - Core Architecture, (Kvm)

hth, Martin.

Jeff Earickson | 3 May 2012 16:18

Re: new "pkgdep" command

All,

If you have mucked with perl on Solaris, beware.  Here is what I got:

(71)> /usr/sbin/pkgdep -v SUNWgzip
Can't locate Sun/Solaris/Utils.pm in  <at> INC ( <at> INC contains:
/opt/perl5/lib/site_perl/5.12.1/sun4-solaris
/opt/perl5/lib/site_perl/5.12.1 /opt/perl5/lib/5.12.1/sun4-solaris
/opt/perl5/lib/5.12.1 /opt/perl5/lib/site_perl/5.10.1
/opt/perl5/lib/site_perl .) at /usr/sbin/pkgdep line 13.
BEGIN failed--compilation aborted at /usr/sbin/pkgdep line 13.

I installed later versions of perl in /opt and Sun's /usr/bin/perl got
overwritten by my version.  I did find Sun/Solaris/Utils.pm in
/usr/perl5/5.8.4/lib, but I haven't figured out how to get this stuff
copied from there to /opt/perl5/lib to get pkgdep to work.  I hope
this doesn't bite me later...

Jeff Earickson
Colby College

On Thu, May 3, 2012 at 5:06 AM, Martin Paul <martin.paul <at> univie.ac.at> wrote:
> Not directly related to PCA, but might be handy: In the README of the new
> revision of the "Install and Patch Utilities Patch" for Solaris 10
> (119254-84, 119255-84) I spotted this:
>
>  5012345 request for tool to list package dependencies
>
> And indeed, the patches include a new script /usr/sbin/pkgdep, which prints
> dependencies of a given package. E.g.:
(Continue reading)

Glenn Satchell | 4 May 2012 01:24
Picon
Picon

Re: new "pkgdep" command

Sun's /usr/bin/perl is a symlink, so restoring that is easy.

ls -l /usr/bin/perl
lrwxrwxrwx   1 root     root          23 Feb 17  2005 /usr/bin/perl -> 
../perl5/5.8.4/bin/perl

There are lots of "system" things now using /usr/bin/perl and expecting 
the Sun modules, etc, to be in place, so you should restore that 
symlink, and use something like /opt/perl/bin/perl for scripts that you 
want to use your version of perl.

BTW pkgdep looks pretty useful.

regards,
-glenn

On 05/04/12 00:18, Jeff Earickson wrote:
> All,
>
> If you have mucked with perl on Solaris, beware.  Here is what I got:
>
> (71)>  /usr/sbin/pkgdep -v SUNWgzip
> Can't locate Sun/Solaris/Utils.pm in  <at> INC ( <at> INC contains:
> /opt/perl5/lib/site_perl/5.12.1/sun4-solaris
> /opt/perl5/lib/site_perl/5.12.1 /opt/perl5/lib/5.12.1/sun4-solaris
> /opt/perl5/lib/5.12.1 /opt/perl5/lib/site_perl/5.10.1
> /opt/perl5/lib/site_perl .) at /usr/sbin/pkgdep line 13.
> BEGIN failed--compilation aborted at /usr/sbin/pkgdep line 13.
>
> I installed later versions of perl in /opt and Sun's /usr/bin/perl got
(Continue reading)

Don O'Malley | 24 May 2012 19:46
Picon
Favicon

Heads-up on patches 147440-16, 147440-17, 147441-16, 147441-17

Hi,

A small number of customers have reported panics in recent days with 
Solaris 10 Kernel patches 147440/1 revs -16 and -17.

The issue appears to be triggered by heavy network stress and also 
appears to be have been caused by the fixfor CR 7054587. Investigations 
are ongoing. It's not yet clear how pervasive the issue is, or under 
what exact circumstances the issue occurs.

117440-15 (SPARC) and 117441-15 (x86) are the last known good 
revisions.  Affected customers should backoff the affected KU's revert 
to these revisions.

I've removed download capability of the affected Kernel patch revisions 
temporarily to avoid more customers being impacted until we can 
determine whether they need to be withdrawn permanently. You will see 
attempts to download these patches via pca failing as a result.

A Sun Alert is in the works and I'll provide an update tomorrow.

We'll also be respinning the Recommended Cluster to revert to rev-15.  
Customers should use the April Solaris 10 CPU instead in the interim.

This issue will be addressed in rev-19 by backing out the offending putback.

Best,
-Don

--

-- 
(Continue reading)

Don O'Malley | 25 May 2012 13:20
Picon
Favicon

Re: Heads-up on patches 147440-16, 147440-17, 147441-16, 147441-17

Hi,

Quick update.

Patches 147440-16, 147440-17, 147441-16 & 147441-17 are in the process 
of being withdrawn from MOS.

I will share the accompanying SunAlert for the issue as soon as I get it.

On 05/24/12 18:46, Don O'Malley wrote:
> Hi,
>
> A small number of customers have reported panics in recent days with 
> Solaris 10 Kernel patches 147440/1 revs -16 and -17.
>
> The issue appears to be triggered by heavy network stress and also 
> appears to be have been caused by the fixfor CR 7054587. 
> Investigations are ongoing. It's not yet clear how pervasive the issue 
> is, or under what exact circumstances the issue occurs.
>
> 117440-15 (SPARC) and 117441-15 (x86) are the last known good revisions.
This should read:
147440-15 (SPARC) and 147441-15 (x86) are the last known good revisions.

(Thanks Thomas!)

Best,
-Don

>   Affected customers should backoff the affected KU's revert to these 
(Continue reading)

Laurent Blume | 25 May 2012 16:45
Favicon

Re: Heads-up on patches 147440-16, 147440-17, 147441-16, 147441-17

Hello Don,

Are they? There *were* unavailable this morning CEST (MOS asked for a  
password with no explanation), but they can be downloaded again now.

Laurent

Don O'Malley <don.omalley <at> oracle.com> a écrit :

> Hi,
>
> Quick update.
>
> Patches 147440-16, 147440-17, 147441-16 & 147441-17 are in the  
> process of being withdrawn from MOS.
>
> I will share the accompanying SunAlert for the issue as soon as I get it.
>
>
> On 05/24/12 18:46, Don O'Malley wrote:
>> Hi,
>>
>> A small number of customers have reported panics in recent days  
>> with Solaris 10 Kernel patches 147440/1 revs -16 and -17.
>>
>> The issue appears to be triggered by heavy network stress and also  
>> appears to be have been caused by the fixfor CR 7054587.  
>> Investigations are ongoing. It's not yet clear how pervasive the  
>> issue is, or under what exact circumstances the issue occurs.
>>
(Continue reading)

Don O'Malley | 28 May 2012 17:35
Picon
Favicon

Re: Heads-up on patches 147440-16, 147440-17, 147441-16, 147441-17

Hi Folks,

These patches should be withdrawn now.

The accompanying SunAlert is 
https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&id=1461565.1

Apologies for any inconvenience.

Best,
-Don

On 05/25/12 15:45, Laurent Blume wrote:
> Hello Don,
>
> Are they? There *were* unavailable this morning CEST (MOS asked for a 
> password with no explanation), but they can be downloaded again now.
>
> Laurent
>
> Don O'Malley <don.omalley <at> oracle.com> a écrit :
>
>> Hi,
>>
>> Quick update.
>>
>> Patches 147440-16, 147440-17, 147441-16 & 147441-17 are in the 
>> process of being withdrawn from MOS.
>>
>> I will share the accompanying SunAlert for the issue as soon as I get 
(Continue reading)

Lee Roth | 30 May 2012 18:25

Linux box as local patch server?

Sadly, our shop has begun a migration away from Sun iron/Solaris to
commodity iron/VMware/Red Hat Linux. <sniff>

I've been asked to plan conversion of as many of the Solaris boxes as
possible to Linux.

Currently, my PCA setup is a "local patch server" on my in-house network
that all of the individual Solaris boxes point to for patching, and the
local patch server is then the sole box that pulls patchdiag.xref and
patches from Oracle.

Question: Can anyone think of a reason that the PCA local patch server
could NOT reside on a Linux server?

Thanks!

Lee

Fred | 30 May 2012 19:08
Picon

Re: Linux box as local patch server?

There is no reason. It will work. A patch proxy is nothing more than a CGI empowered apache server. The underlying OS of the proxy plays no part what patches the proxy downloads and provides to it's clients. You could therefore make your patch proxy the first Linux host on the road to migration -- ironic though that may be.

On a completely seperate note, it could be said that VMWare licenses + RHEL suport + commodity iron is about the same price as Oracle iron + OVM + OEL (the latter two of which have no cost when deployed on Oracle hardware). But everyone's mileage varies when it comes to what deals they cut with their vendors.

Fred

On Wed, May 30, 2012 at 12:25 PM, Lee Roth <patch20 <at> easy48.com> wrote:
Sadly, our shop has begun a migration away from Sun iron/Solaris to
commodity iron/VMware/Red Hat Linux. <sniff>

I've been asked to plan conversion of as many of the Solaris boxes as
possible to Linux.

Currently, my PCA setup is a "local patch server" on my in-house network
that all of the individual Solaris boxes point to for patching, and the
local patch server is then the sole box that pulls patchdiag.xref and
patches from Oracle.

Question: Can anyone think of a reason that the PCA local patch server
could NOT reside on a Linux server?

Thanks!

Lee






--
Fred Chagnon
fchagnon <at> gmail.com
Martin Paul | 31 May 2012 08:46
Picon
Picon
Favicon

Re: Linux box as local patch server?

Lee Roth wrote:
> Question: Can anyone think of a reason that the PCA local patch server
> could NOT reside on a Linux server?

No, that should be fine. While I'm not using such a setup here, I've had reports 
from people who run a PCA local caching proxy on Linux machines successfully.

Martin.


Gmane