re: building identifiers
by Adam Stokes
Adam Stokes wrote:
> So I am trying to figure out the best way to create identifiers on a
> per-machine basis.
>
> Identifier for use in the file name? I think your right, RHN ID if
> it's there, perhaps FQDN if it's there and not
>"localhost.localdomain",
> and then some MD5 hash if nothing else. For RHN ID, tack something
> like "RHNID" to the front of the ID so that we can distinguish it
> from
> an MD5 hash or whatever new scheme we come up with in the
> future.
Yea I think I'll add the checks for at least RHNID and if hostname is
not localhost we'll use that. Good idea.
> Additionally, I've been thinking that it would be good if we attach
> this kind of information to core dumps (like RHN ID, and hostname, and
> ticket #, etc), into a file (or files) in the tarball that contains
> the
> core dump so that we aren't limited to just the information we can fit
> in a file name. Thoughts?
Yea I was thinking about this too, perhaps some kind of metadeta that
has that type of information in it. So far I only have one plugin
written for fastback but netdump is next and I believe just one file
that has the above information in it is necessary.
I'll know more about what I'm saying when I get some more code
written :)
--
__ __
.-----.| |_.-----.| |--.-----.-----.
|__ --|| _| _ || <| -__|__ --|
|_____||____|_____||__|__|_____|_____|
.----------------------------------.
( ajs(a)redhat.com || 919.754.4187 )
`----------------------------------'
15 years, 3 months
building identifiers
by Adam Stokes
So I am trying to figure out the best way to create identifiers on a
per-machine basis.
So far I'm given the option to define a name in the configuration file
and then prompting users for a case number or generating a md5 has if
one doesn't exist. Building the identifier off of a rhn-id doesn't
seem full proof as some systems never even connect to RHN and fedora
systems do not require any sort of registration when getting updates
from the repo's.
The way im doing it so far seems pretty straightforward, however, I'd
like to get some feedback on if we should store a case number/md5 hash
in the configuration file for each system to help us further identify
systems based on incident reports. With this we can provide an option
to override that case number at any time with a simple command line
switch.
Any thoughts?
--
( adam stokes ) || ( adam.stokes(a)gmail.com )
15 years, 4 months
Re: fastback on fedorahosted
by Gavin Romig-Koch
I've moved (with lots of changes) the wiki pages from tools.gss to
fedorahosted. I've tried to improve the explanation of the
goals/design. Please read it over, comment, question, retch, improve.
-gavin...
15 years, 4 months