[rhn-users] Re: rhnsd daemon update

[Jesse Becker]

On Wed, Jan 19, 2005 at 12:27:10PM -0500, nathan r. hruby wrote:
> >Right, now will this apply all updates (critical and feature/bugfix
> >releases), or just critical updates?
> >
> 
> I would think all of them.  We don't use the auto update feature, so I
> couldn't tell you.

Noted.  I've observed variable behaviour:  sometimes everything is applied,
sometimes only critical patches.  I do not have good test cases, and haven't
been able to isolate anything specific.  I do know that sometimes rhn_check
and up2date will want to do different things.

> I asked before, why do you think that rhnsd dies/sticks?  What behavior

I think it dies/sticks because I've seen it get stuck, and machines fall out
of date, even when rhnsd is nominally running.  I've also seen the process
vanish, and concluded that it dying on it's own is the most likely cause (but
not the only one).  Usually it just gets stuck though.

> have you seen or evidence that you have supports it is rhnsd that is the
> problem?  

As I mentioned, I don't have a good test case.  It's a recurring, not
repeateable, problem.  One thing that usually "unsticks" things is to bounce
rhnsd.  It is entirely possible that I'm just missing something completely in
the configuration...

> Have you submitted a bug to bugzilla.redhat.com about it (if so
> what's the bug number)?  Have you opened a support ticket with Red Hat
> Support?

No, and no, respectively.  I don't have enough information (yet) to file
a proper bug report (i.e. good error messages, steps to reproduce, etc).
When/if I ever do, I will certainly file something.

> I say this because I have many dozens of boxes that run rhnsd and have
> never had a problem with it.  OTOH, the RHN website (both hosted and
> satellite) has some serious caching issues (or something, I have not had
> the time to track it down and have been poo-poo'ed by support in the past
> about it) that causes Erratta to be incorrectly reported for a machine
> until such time as you do a package refresh for said machine.  Perhaps
> this is what you attribute to rhnsd being "stuck"?

That is possible.  I've noticed some inconsistency between the RHN pages and
actual state of the machine.  In your cases, have you found running
'rhn_check refresh' to be handled the same as scheduling a profile update from
RHN?  In case it matters, I'm just using a Proxy, not a full-blown satellite.

-- 
Jesse Becker
GPG-fingerprint: BD00 7AA4 4483 AFCC 82D0  2720 0083 0931 9A2B 06A2
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/rhn-users/attachments/20050119/c617f0b8/attachment.sig>