Find the answer to your Linux question:
Results 1 to 6 of 6

Thread: Red Hat Ent 5

Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1

    Red Hat Ent 5

    Could really use some hekp with a Proliant 380 that has the above version of linux. Not really profficient with linux and am I trying to connect a this server to a NetApp for storage.

    I can ifconfig eth1 & eth2 and both are up with different IP's and netmask, they can ping each other, but I can not get eth1 to ping the NetApp, both are in the same subnet. Perhaps I am wrong, but it may be my routing table. when I do a route, I eventually get 2 line of errors and then the gateway will display. The errors state /etc/host.conf:line 3 (and 4):bad command IP FQDN.

    I have vi the host.conf and deleted these entries, but they still come up. Can anybody direct me to the directory as I'm sure there is another file that still has these entries.

    Further, there are 4 systems on this subnet, 1 NetApp, 2 Win 03 and this linux. All can ping the linux box, but it can't ping out.

  2. #2
    Just as needed if this was Windows, what is the NIC config and routing table? And what is the NIC config on the NetApp? (I assume you're using an iSCSI LUN off the NetApp, so beyond NIC config, you'll be getting into iSCSI initiator config.)

    # route
    Kernel IP routing table
    Destination     Gateway         Genmask         Flags Metric Ref    Use Iface     *        U     0      0        0 vmnet1    *        U     0      0        0 vmnet8
    10.X.X.X      *        U     0      0        0 eth0
    link-local      *          U     0      0        0 eth0
    loopback        *            U     0      0        0 lo
    default         10.X.X.X         UG    0      0        0 eth0
    # ifconfig 
    eth0      Link encap:Ethernet  HWaddr 00:C0:9F:X:X:X  
              inet addr:10.X.X.235  Bcast:10.X.X.255  Mask:
              inet6 addr: fe80::2c0:9fff:fe29:d340/64 Scope:Link
              RX packets:4666556 errors:0 dropped:0 overruns:0 frame:0
              TX packets:2462892 errors:0 dropped:0 overruns:0 carrier:0
              collisions:0 txqueuelen:100 
              RX bytes:2770393299 (2642.0 Mb)  TX bytes:198944792 (189.7 Mb)
              Base address:0xecc0 Memory:fe100000-fe120000
    Ping by name = *name resolution*, then ARP, then send ICMP. Ping by IP = Send ARP, then ICMP.

    Two NIC's on the same subnet is generally a bad idea - only name resolution will allow other hosts to tell them apart. You say "they can ping each other", so they are either on the same network (assuming you are forcing traffic to go out and back in, otherwise, you are just getting a response from the local NIC), or you have a loop condition (usually not good.) Bonding them into 1 interface may be an option.

    You can also check to see if firewall (iptables) is blocking on the RH box as well as if SELinux is installed/enabled.

    * Edit: Routing is not involved when everything's on one subnet - assuming the network config is correct/valid.

  3. #3
    the proliant has 2 nics (eth1 & eth2) with each config for 2 differnet subnets, yet the only thing they can ping is each other. Eth1 is on the same subnet as the NetApp and 2 windows servers. Eth1 can't ping anyone of them by IP or name. Eth2 doesn't matter at this point other to state it is config and that it can ping eth1.

    Yes window servers using iscsi and now trying to connect linux server to it. I turned off iptables so that firewall would be disabled. Have not tried arp or arp -a as i seen online yet. Tried host.conf with IP and FQDN and w/o, but everything I have tried gives me destination unreachable.

  4. $spacer_open
  5. #4
    You have provided no info to work with.

    Your main config files are /etc/hosts, /etc/nsswitch.conf, and /etc/resolv.conf. But all of them deal with name resolution. If you are using IP's, there is no name resolution and these files all mean nothing.

    Is the NIC "up", does it have link, is the speed set/negotiating correctly, is the switch port set for the correct VLAN, etc.

    Unless you have a traceroute showing multiple hops from NIC2 to NIC1, you are probably just getting a response from the local NIC and not going out through the network and back. If the NIC's are on different subnets, you *must* have something routing between the subnets - even if they're just VLAN's on the same switch.

    As always, you have packet capture to fall back on.

  6. #5
    Ok, hows this? Eth1 and the NetApp is at and both can not ping(IP/Name)/arp and icmp each other.

    Disabled firewall on eth1 server iptable, can not realistically do on NetApp.

  7. #6
    Those are not on the same subnet...

    Based on your netmask, your subnets are: - -

    The two IP's you listed are in different ranges.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts