[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [cpx] [CPX] - serious issue in handling of the users
- Subject: Re: [cpx] [CPX] - serious issue in handling of the users
- From: ADNET Ghislain <gadnet@xxxxxxxxxx>
- Date: Wed, 30 Mar 2005 11:03:26 +0200
So if I may restate:
All end user accounts with a dot in their usernames suddenly reverted
their domain to the server hostname.
Is this correct?
 
mostly, all end users of the DOMAIN reverted to the hostname, i just 
noticed they have  2 users without DOT in the name (that also reverted) 
but they almost all have a dot in them.
Was there any trigger you can think of? Did you perform any kind of
CP operation (add a domain, etc.) when this happened? Any additional
information you can give, including confirmation of the above will
help us pinpoint the problem.
 
not at all, no domains were added. I know support created users and 
changed webmail/spam/virus settings for some and probably deleted users 
too.  They touched the server just the same day the customer discovered 
the new issue. On this server only one domain admin has the access, and 
my customer has server admin access to manage his domain admins (who do 
not have access apart from one ).
 The problem is that i have no log of what is done in the cpx (HINT 
have a debug mode that log things somewhere that you can trigger as a 
severadmin ;)  ). My customer say he has not done anything those last 
days with it. if i look in /home i do not see new users for those days 
so i bet no websites have been created and i destroyed the test users 
support forgot to destroy when they looked at it (or i suppose as i do 
not know a way in the cpx to create users in the wheel group and the 
customer doesn't even know that this group exist).
In the user.log i have no entry between the 23 and 28 march and the two 
are test users created by support. So it seems the customers does not 
created anything in this  period.
In cpx.log i got a lot of error like this:
Use of uninitialized value in -f at 
/usr/local/lib/perl5/vendor_perl/5.8.4/VSAP/Server.pm line 419.
 I just noticed those two log ?? is this new i do not see them in other 
VPSes...
I have destroyed the cpx.conf  since to be sure there was no corrupted 
things in it so cpx reconstructed his clean own one. But i still got the 
problems of users moving back and forth as i change them.
 Since then  i tried to move again users but  i got the problem that 
you move one user and one other get moved back to the hostname so it's 
kind of impossible to do it with reliability... i do not found any 8bit 
chars in the fullusername field of the password that can trigger a 
problem, i just removed any '&' and "'" in this field just to be sure it 
does not trigger some problem also.
i am kind of lost here...
Is anyone else seeing something like this with the same conditions?
Scott
 
--
AQUEOS - Service Informatique
1, Rue Albert Einstein
77420 Champs sur marne
 	
Service technique :  support@xxxxxxxxxx
Service commercial :  commercial@xxxxxxxxxx
Tel : 06.63.79.27.38 /  01.64.02.99.37
======================================================================
This is <cpx@xxxxxxxxxxxxx>      <http://www.groupmail.org/lists/cpx/>
Before posting a question, please search the archives (see above URL).
Home |
Main Index |
Thread Index