Okay, I just purchased a Cobalt server to play with and learn things.

So, one of the first questions I have is this: Why are there two Ethernet ports on the box? What is the point?

Why run two different cables into the box? Is it for redundancy? If so, do you also assign two different IP's to each domain you set up on the box? One primary and one secondary??

Can someone answer this for me, or point me to a site with detailed information.....

Thanks!

Hi,

On our raq 4's the reason we use two ethernet ports is simply for backing up, this can be done with out being charged for internal data if you have a backup server of course. An yes you will need 2 seperate ip's

Hope this helps

So, the second port is used to transfer information to a back-up server, therefore saving on bandwidth costs....is that correct?

It's a Cobalt RaQ4r I bought....

Correct,

Well that is how it works in our data center ;)

You could also use it to connect to dedicated database server if being used as an web application server...

or a firewall (bit overkill for a 4r, but good application for a 3i)

Brandon

Hey All,

The 2 NIC cards have many purposes, but when Cobalt added them to the initial 3i they had their StaQware fault tolerence software in mind (says my contact at cobaltnet). However StaQware was never very sucessful, it had lots of little problems. I have never had a problem or heard of any major problems involving staQware, but when there is a problem I have been told it is hard to fix due to the complexity of StaQware..

For thoes who dont know what StaQware is it is a small PKG that you can install on 2x 3i's or 4i's, it is still available from cobalt downloads if you have a spare pair of raqs.. You set it up to have one server as your primary and the other as your seccondary server, you connect them both to the public internet or intranet using the 1st NIC and with the 2nd NIC you use a crossover cable or a VLAN between the two servers to form a 100Mbs Full Duplex link.

The primary server you use as you use a normal cobalt, adding virual sites, intalling packages etc while the other sits behind it invisable. The private network connection between the two servers has two jobs the first being to let the servers continually monitor each other to make sure they are on and to make sure all services are working properly, the other job is to continually syncronise the data on both servers. In the event of a fault the seccondary server will determine how serious the problem is (eg if the webserver crashes make sure it starts up properly) and if it determines that the primary server will not be able to recover the problem it will start the change over procedure. This will disable the primary server's IP addresses and enable them on itself. This takes 2-10mins to do. Once this is done the backup server should look just like the primary server once looked.
You can then restore the old cobalt and put it back with the other one, it will then be there as the seccondary server should somthing happen to it.

Overall it is a great system for those of us who need a higher level of fault tolerance. If you have a spare pair I would reccomend playing around with this system, its really interesting... :)


As Marts said the the second NIC is great for backup systems. In our case we have a robot script that will automatically retreive backup files from our dedicated server clients' servers on a regular basis, the script copies the data to the backup server and writes tapes either daily weekly or monthly depending on the clients needs. We just use a cheap 100meg switch to sit behind the network with all servers using their seccond network card and a private ip address connected, this means that all have secure access to our backup server natuarlly with unmetered traffic.


I have also setup my machine as a network gateway and firewall which works quite well.. I have a RPM for a firewall that works for the raq using the 2 NICs if your interested.

Hope this helps

Regards,
Samuel