|
Ted Pavlic wrote:
>
> I have a feeling that it has something to do with the NAT rather than LVS.
>
> I use LVS to load balance about 1024 VIPs onto 4 real servers using VS-DR. I
> have not had any problems with only 4096 file handles. I'm very sure that I
> doubt I use even 1000 file handles under normal operation.
That's the second response along those lines that I've heard regarding
using VS-DR over VS-NAT. The reason I picked the VS-NAT method was
because I had seen in a couple of different places in the documentation
that VS-NAT was really the only fully-functional method at the moment.
Maybe the LinuxVirtualServer.org website is outdated, then? VS-DR would
actually fit our current network model better anyway.
> Could you include an example output from an ipvsadm -L ? About how many
> active connections do you normally get to each server? About how many of
> active connections total using the aggregate of all of the servers? How
> about inactive connections?
I could give an example output from ipvsadm -L, but it doesn't look like
anything's running right now.... (One of the problems with diagnosing
the problems here is that the client keeps bringing machines up and down
behind the LVS boxes so sometimes I can't tell if the LVS is not working
or their server has just crashed...)
Right now, we don't really have any hard data for connections since
we're still setting up the boxes, and as I said, the client keeps taking
services down for reasons they don't explain to me.
Also, I have a real problem in that we have confidentiality agreements
with our client that would prevent me from giving out too much
information. I can probably "sanitize" the output and config files and
whatnot, but I'm not sure they would be entirely useful at that point...
> Now I usually don't do much kernel or filesystem development, but the inode
> problem you mention makes me think even more that it isn't an LVS problem.
That was one of my big questions. If other people are doing the same
thing with LVS and not having these problems, then we're obviously doing
something different where different==wrong. If other people see same
problems, then it's just something we have to keep an eye on.
> That's just my two cents. From your description of what you're doing, I
> really don't think that you're doing anything too exotic.
Ahhh nuts... :)
> Perhaps you could send us some of your configuration files? Perhaps the
> output of some of the RedHat-HA utilities? Give us a better model of your
> system. Maybe someone will come up with some other ideas.
If I can convince someone there to put some services back online, I'll
try to get some more detailed config data. I'm also thinking about
looking at Ultramonkey and see if a change in the toolset we're using
clears up the problems.
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
With Microsoft products, failure is not Derek Glidden
an option - it's a standard component. http://3dlinux.org/
Choose your life. Choose your http://www.tbcpc.org/
future. Choose Linux. http://www.illusionary.com/
|
