Le lundi 10 janvier 2011, Sylvain Beucler a écrit :
> chkrootkit has been triggering a lot of false positive (like SQL
> injection monitor ;)).  In addition it can only be run securely on a
> trusted box, because an attacker installing a rootkit can patch or
> disable chkrootkit on a compromised system.  I don't remember if it
> was installed at Gna! but we didn't keep it at Savannah for these
> reasons.
> 
> (Jame Villate found the Savannah rootkit in 2003 btw, not Vincent, but
> that's an old story now)

I don't remember exactly but I'm quite sure Vincent was at least responsible 
for installing chkrootkit. (I'm sure, anyway, Jaime know I don't mean to 
underestimate it's involvement; it's no secret, for instance, that without him, 
there would have been no CERN contribution at all)

I'm not saying chkrootkit is the ultimate answer. I'm just saying I'm finding 
it very hard to track down current security measures while there were plenty of 
little ones (no ultimate weapon, though, that's correct) at start that 
diseappear with no obvious replacement. For instance, is there any files 
integrity checks done?
  
> /root/TODO is basically my personal roadmap when I was working mostly
> alone on upgrading the infra.  I didn't think much about it but it was
> more convenient to use than a dozen of separate tracker items in that
> particular case.  We use trackers for a number of things but we also
> know when not to use them.

All projects I know of make use of trackers not because it is fast. There is an 
obvious overhead. But the point is to communicate and allow other to keep 
tracks about what is going on.
The very idea of "personal roadmap" is quite annoying. Even if I understand 
well how it is practical.

 
> /root/.ssh/authorized_keys is managed automatically in the host
> (Bearstech procedure).  The guests are managed manually because nobody
> took the time to install the script.  Note that the script does not
> automate the task for security reasons, instead it suggests what keys
> to add.  Mixing Savane users and system users doesn't sound good, so
> better leave the backup key as-is.

How does it not sound good? As you stated, there is no security issue about it.
So it is because nobody took the time to install it or is it because it is not 
a so good idea?

You brought up the perfect example of what I'm trying to talk about:
Everything looks that way, things that were are goner but there's no clear 
explanation: maybe it is a matter of time; but maybe it is because one (or 
more) thought it not so great.
So, when you take a look once in a while, you notice something no longer 
exists, what are you supposed to do? Take the time to fix it? Assuming that 
it's gone for good reasons? You send a mail each time, bothering people about 
their choices? Frankly, first reaction is to move on and do something else, 
elsewhere.
All over, it is the same thing. You have to second-guess everything. Or to 
bother people asking them to justify whether there is a good reason for some 
removal or if it just because they lack time, while at the same time you devote 
no time at all. It's just annoying.

> Your last remark shows you're aware of the tone of your mail :/

It's not a matter of tone. I cannot think of a nice way to tell people that we 
think things should be done differently, especially when you are not doing 
anything at all. And, anyway, by nature, mails are harsh, even if you put 
smileys everywhere.

I thought useful to say that I'm not questioning what is done. From what I see, 
it is very good. I'm just talking about what is not done and how I do not think 
it can be addressed in the current state of thing. Unwritten policies, 
unexplained choices, I dont see how it can work if we want more than one person 
in charge at a time.


-- 
Mathieu Roy

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
Project mailing list
[email protected]
https://mail.gna.org/listinfo/project

Reply via email to