Hello,

I'm interested in putting together a gentoo mangement node subsystem to aid in 
multi-gentoo administration.

The immediate target is checking for security updates.

For this, I've come across a problem regarding root handling which I've opened 
a bug for here:
  https://bugs.gentoo.org/show_bug.cgi?id=129054

After sorting this issue, I was able to successfully test a remote machine for 
glsa on the local node, by transferring its VDB to the local machine, and 
also find out which packages needs updating. 
Taken into account that the remote machines are backends which by security 
requirements do not have any access to the internet, this saves the manual 
work of downloading the packages and placing them of the servers - and this 
worlds quite well, since I can emerge -f for a remote machine.

Another detail I need to sort out - does portage use ${ROOT}/etc/ or 
just /etc ? Since I would need to store the useflags and customization for 
each remote host, it is a requirement.

I've came across another problem as well.
importing portage will go through its initialization, checking VDB and sorts, 
so that if I'd like in to switch a VDB I'll have to re-import portage. Since 
I do not know of a way to re-import portage in python, I've worked around 
this by changing ROOT env and then call portage via os.system calls.
On my humble amd 2600 machine this yields a 8 seconds average per VDB change.

Is there something I can do beside replicating the init code into a function, 
change root and recall the init function again?

Is there a better way of doing this?


-- 
Eldad Zack <[EMAIL PROTECTED]>
Key/Fingerprint at pgp.mit.edu, ID 0x96EA0A93

Attachment: pgpQnxVpwQk4c.pgp
Description: PGP signature

Reply via email to