I have a server that I have inherited from someone who left our company
and he built it with Debian 3.1 on the root server and all guest vservers.
The kernel used was 2.6.8-vserver. I'm going to ask some dumb questions
here because I'm mostly familiar with RedHat and Solaris, more than
Hi Herbert --
Thanks for the reply. You're making me feel more comfy about this whole
thing. :)
Interesting point about putting RedHat on there for the host OS. We're
starting to move all of our corporate servers that are Debian to a
supported operating system (i.e. either by Dell,
I totally agree that it just sounds bizarre. What you said is basically
the only thoughts I could come up with as well. When I get a chance I'll
perhaps pursue the nameserver problem idea.
Kathy
that sounds indeed very weird ... my best guess here would
be that you have some kind of local
Been looking at this problem for 2 days now and can't figure it out --
hoping for some ideas as to what to look at.
I just created a new Debian guest. Called it wwwint. My Debian
distribution is on another machine called bermuda. I created the vserver
by the following command:
vserver
Well, figures after I make a cry for help, I got it to start behaving.
AFter probably the 10th reinstall of the guest, I started getting
suspicious
of the name I was using. This may have no bearing on what the real
problem was but I was calling this guest wwwint and I already have a
I have a vserver Debian guest in which I have been creating lots of
IP aliases and yesterday it gave me the error that there is a limit of 16.
Is there a way to increase that limit in the kernel? Does this limit of
16 also apply to the root server? My concern is that I have 3 separate
Nope. Once an alias is created, however which way, the system comes up
with a 16 alias max message. Don't recall the actual message. The
vserver would not restart and I had to remove the interfaces and reboot
the entire server.
Kathy
Does it help to add ip's with the ip add ...
Thanks for the info Bruno. Helps a lot.
Kathy
The kernel patch has a maximum of IP addresses per network context defined.
Context-association of IP addresses is not optimized, thus many addresses
cause some overhead when checking if incoming packets may be delivered to a
guest. (check
That would be a good thing. :-) But I still have my other question
that I'm concerned about. If I have two guests with 10 aliases each, such
that my root server sees 20 aliases, that still would send the entire
server beyond the current 16 limit, correct? I guess it's sort of a dumb
Has any one seen a problem similar to what I'm having? I have a webserver
that is running 5 different vservers, one being the root server. They
each have individual IP addresses as defined in
/etc/vservers/vserver/interfaces. I'm seeing the problem where if I
start postfix in the root
What does root server mean? The physical host, containing all the
guests?
Yes. The main system running the host OS. root server is what
vserver-stat reports it back as.
If the above is true, that is expected behaviour. You have to bind
services on the host to the correct IP addresses,
Thanks, Nicolas, for the reply. I have just now and gone and bound smtp
to their specific IP addresses and restarted postfix on all. I can see
your point about not wanting to run anything on the root server. Despite
binding all postfixes, including the root server's, the root server
I posted this before but not sure it actually went to the list, being my
first post.
I am having troubles with a vserver (named www) in that when I add a 6th
IP alias to it, it will not create the interface when the entire system is
rebooted. And until I remove that 6th definition and
Hello --
I've looked around in the archives but couldn't find anything similar to
this situation. I have a server that is running Debian (sarge, kernel
2.6.11-vserver) and runs 4 vservers. Two of the vservers are webservers,
which run Apache2. I am having troubles with the vserver named
14 matches
Mail list logo