Why do some patches report The patch XXX makes the kernel unstable

Some patches report on installation with smpatch

The patch 119252-13 makes the kernel unstable, is obsolete or is forbidden by policy.

The patch 119280-07 makes the kernel unstable, is obsolete or is forbidden by policy.

Theres nothing particularly noteworthy mentioned about the patches in their README files.

Its a different message than we get from single user patches or interactive patches.

They don't sound like the kind of thing that would make the kernel unstable.

And in that case, surely they should be single user reboot immediately.

If their obsolete then why are they turning up in smpatch analyze.

And if their against policy, what policy is it?

I havent noticed any particular problems from forcing them in with a manual smpatch -i

I just wondered whats special about them.

Its a bit annoying having to add them manually...

[914 byte] By [robert.cohen] at [2007-11-26 10:35:15]
# 1

Hi.

the first patch has special install instructions, so smpatch should not be installing it.

The second patch, I'm not sure about. I can't see anything unusual about this in the patch description. Could you post the output from 'smpatch get' - there may be something in your patchpro.install.types values which is causing smpatch to flag this patch

ForumModerator at 2007-7-7 2:45:32 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 2

Its all at the default values.

talwood# smpatch get

patchpro.backout.directory-""

patchpro.baseline.directory-/var/sadm/spool

patchpro.download.directory-/var/sadm/spool

patchpro.install.types -rebootafter:reconfigafter:standard

patchpro.patch.sourcehttp://bondi.anu.edu.au:3816/solarishttps://getupdates1.sun.com/

patchpro.patchset-current

patchpro.proxy.host -""

patchpro.proxy.passwd********

patchpro.proxy.port -8080

patchpro.proxy.user -""

robertcohen at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 3

Robert,

I've done a little bit of testing and digging around here and I think i've discovered a possible cause for the patches to react in this way. At this time though I do not know whether this is by design or if it's a bug (or even if my suspicions as to the cause are correct). As such i'll consult with other people internally and post an update when I have anything further.

ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 4

I have the same issue with these patches (both the x86 and sparc forms).

BTW rather than this alarming message it would be better if smpatch simply reported that there were special requirements and the patch had to be installed manually. Of course then you have to know to unpack the jar file, read the README and follow the instructions.

As for the special install instructions for 119252-13, I'm not a sys-admin and I haven't got a clue what the mini-root is or how to patch it or use it. :(

davidholmes at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 5

Dear David,

Agreed the error could be worded in a less alarming manner.

The miniroot is the reduced size root filesystem generally used when booting or installing a client from a server over the network - jumpstart etc and the old fashioned diskless clients situation sometimes.

We will keep you (both) posted on our findings re the error messages.

ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 6

The cause appears to be a metadata tag that has been incorrectly associated with these patches which would indicate something that is not truely a Solaris patch such as firmware images. The patches themselves are fine.

This is currently being investigated and when we receive an update we will post to this topic.

ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 7
A quick update - two bugs have been raised (one for the patches, and one for the Sun Update toolset).Mod.
ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 8

I note that the latest versions of 119253/62 still have this problem.

UpdateManager 1.0.9 flags it as a "download only" patch, but still allows you to select it to install and then the install fails. Does "download only" mean "updatemanager will download this but you need to follow the manual installation instructions in the readme"? If so then the "icon legend" should mention this.

How is the end-user (not a trained sys-admin) supposed to understand how to deal with these sorts of patches ? The information in the readme just leaves me wondering what on earth it is I'm supposed to do. Taking care of this is what I expect updatemanager to do!

davidholmes at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 9
davidholmes - this should not happen, unless smpatch has been specifically configured to allow these patches to be installed. Could you post the output from:$ /usr/sbin/smpatch get patchpro.install.types
ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 10
root hypert /var/sadm/spool # /usr/sbin/smpatch get patchpro.install.typesrebootafter:reconfigafter:standard
davidholmes at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...
# 11
What is the error given when the patch install fails?
ForumModerator at 2007-7-7 2:45:33 > top of Java-index,Administration Tools,Sun Update Connection-System...