LPAR Networking: Can't ping

PPPP · July 17, 2012, 10:11am

So I have created an LPAR. I think I have configured everything properly but I can't ping the network. The LPAR is on a VLAN. Other LPARs on the same VLAN works. Here's some info on the LPAR:

# entstat -d ent1
--------------------------------------------------
ETERNET STATISTICS (en1) :
Device Type: Virtual I/O Ethernet Adapter (l-lan)
Hardware Address: 16:72:10:00:90:03
Elapsed Time: 0 days 1 hours 11 minutes 30 seconds

[snip]

# netstat -rn
Routing tables
Destination        Gateway           Flags   Refs     Use  If   Exp  Groups

Route Tree for Protocol Family 2 (Internet):
default            10.1.40.1         UG        1        44 en1      -      -
10.1.40.0          10.1.40.41        UHSb      0         0 en1      -      -   =>
10.1.40/24         10.1.40.41        U         4       932 en1      -      -
10.1.40.41         127.0.0.1         UGHS      0         1 lo0      -      -
10.1.40.255        10.1.40.41        UHSb      0         0 en1      -      -
127/8              127.0.0.1         U         4        70 lo0      -      -

Route Tree for Protocol Family 24 (Internet v6):
::1                ::1               UH        0         0 lo0      -      -

# ifconfig -a
en1: flags=1e080863,480<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST,GROUPRT,64BIT,CHECKSUM_OFFLOAD(ACTIVE),CHAIN>
        inet 10.1.40.41 netmask 0xffffff00 broadcast 10.1.40.255
         tcp_sendspace 262144 tcp_recvspace 262144 rfc1323 1
lo0: flags=e08084b<UP,BROADCAST,LOOPBACK,RUNNING,SIMPLEX,MULTICAST,GROUPRT,64BIT>
        inet 127.0.0.1 netmask 0xff000000 broadcast 127.255.255.255
        inet6 ::1/0
         tcp_sendspace 131072 tcp_recvspace 131072 rfc1323 1

# netstat -in
Name  Mtu   Network     Address           ZoneID    Ipkts Ierrs    Opkts Oerrs  Coll
en1   1500  link#2      16.72.10.0.90.3             17119     0      848     0     0
en1   1500  10.1.40     10.1.40.41                  17119     0      848     0     0
lo0   16896 link#1                                     70     0      916     0     0
lo0   16896 127         127.0.0.1                      70     0      916     0     0
lo0   16896 ::1                                0       70     0      916     0     0

I can't figure out why I can't ping anything other than 127.0.0.1...not even the default gateway @ 10.1.40.1. Any idea?

# ping 10.1.40.1
PING 10.1.40.1: (10.1.40.1): 56 data bytes
^C
----10.1.40.1 PING Statistics----
3 packets transmitted, 0 packets received, 100% packet loss

# lsdev -Ccadapter
ent0   Available 01-08 10/100/1000 Base-TX PCI-X Adapter (14106902)
ent1   Available       Virtual I/O Ethernet Adapter (l-lan)
ide0   Available 03-08 ATA/IDE Controller Device
usbhc0 Available 02-08 USB Host Controller (33103500)
usbhc1 Available 02-09 USB Host Controller (33103500)
vsa0   Available       LPAR Virtual Serial Adapter
vscsi0 Available       Virtual SCSI Client Adapter

# lsdev -Ccif
en0 Defined   01-08 Standard Ethernet Network Interface
en1 Available       Standard Ethernet Network Interface
et0 Defined   01-08 IEEE 802.3 Ethernet Network Interface
et1 Defined         IEEE 802.3 Ethernet Network Interface
lo0 Available       Loopback Network Interface

# lsattr -El en1
alias4                      IPv4 Alias including Subnet Mask           True
alias6                      IPv6 Alias including Prefix Length         True
arp           on            Address Resolution Protocol (ARP)          True
authority                   Authorized Users                           True
broadcast                   Broadcast Address                          True
mtu           1500          Maximum IP Packet Size for This Device     True
netaddr       10.1.40.41    Internet Address                           True
netaddr6                    IPv6 Internet Address                      True
netmask       255.255.255.0 Subnet Mask                                True
prefixlen                   Prefix Length for IPv6 Internet Address    True
remmtu        576           Maximum IP Packet Size for REMOTE Networks True
rfc1323                     Enable/Disable TCP RFC 1323 Window Scaling True
security      none          Security Level                             True
state         up            Current Interface Status                   True
tcp_mssdflt                 Set TCP Maximum Segment Size               True
tcp_nodelay                 Enable/Disable TCP_NODELAY Option          True
tcp_recvspace               Set Socket Buffer Space for Receiving      True
tcp_sendspace               Set Socket Buffer Space for Sending        True

zaxxon · July 17, 2012, 11:56am

Is 10.1.40.41 an external network or is it virtual and does only exist in on your managed system?

PPPP · July 17, 2012, 11:57am

10.1.40.41 is the IP of the LPAR I'm trying to setup. It only exists in my private network.

zaxxon · July 17, 2012, 12:29pm

But you ping your default GW 10.1.40.1 in your example. Is there any such?

PPPP · July 17, 2012, 12:30pm

I can't ping even the default GW. You will notice it says 100% packet loss.

zaxxon · July 17, 2012, 12:53pm

Maybe you are in the wrong VLAN. Check the VLAN-ID of your adapter.

PPPP · July 17, 2012, 12:56pm

How do I check that on the command line?

In the HMC, I was able to graphically confirm that the LPAR is on the same VLAN as another LPAR that is connected to the network, but on a different subnet.

Partitions VLAN 1* 
AIX DEV  
AIX_CSS  <--- Working LPAR
AIX_QA  
AIX_SA  <--- My LPAR
QA-AIX61  
SLES9  
VIOServer   
  
* Indicates the VLAN is bridged

zaxxon · July 17, 2012, 1:27pm

You need a gateway or an adapter into that other subnet. So is 10.1.40.1 a gateway? Is it up and running?

PPPP · July 17, 2012, 1:34pm

Yes, 10.1.40.1 is a gateway and it's up and running. My other systems using that as the default GW is up.

zaxxon · July 17, 2012, 3:13pm

Is the gateway 10.1.40.1 external ie. not an LPAR but a router? If yes, you might be missing a SEA (Shared Ethernet Adapter) to bridge the VLAN to the physical LAN. Maybe compare with the LPARs that are working, if they use a SEA.
In the output of your lsdev -Ccadapter I see no SEA listed. That could be it.

PPPP · July 17, 2012, 3:42pm

I'm 99% sure 10.1.40.1 is external because my non-LPAR systems use that gateway as well.

How do I know if an LPAR is using a SEA? I know I have a SEA because on the VIO Server...

$ lsmap -all -net
SVEA    Physloc
----- ----------------------------------------
ent2  U9110.51A.06A027D-V1-C3-T1

SEA                     ent3
Backing device          ent0
Physloc                 U788C.001.AAAA6218-P1-T1

On the LPAR that's working....

# lsdev -Ccadapter
ent0   Available         Virtual I/O Ethernet Adapter (l-lan)
ide0   Defined     01-08 ATA/IDE Controller Device
vsa0   Available         LPAR Virtual Serial Adapter
vscsi0 Available         Virtual SCSI Client Adapter

And on the LPAR that's not working:

# lsdev -Ccadapter
ent0   Available 01-08 10/100/1000 Base-TX PCI-X Adapter (14106902)
ent1   Available       Virtual I/O Ethernet Adapter (l-lan)
ide0   Available 03-08 ATA/IDE Controller Device
usbhc0 Available 02-08 USB Host Controller (33103500)
usbhc1 Available 02-09 USB Host Controller (33103500)
vsa0   Available       LPAR Virtual Serial Adapter
vscsi0 Available       Virtual SCSI Client Adapter

zaxxon · July 17, 2012, 3:51pm

Right, SEA is only visible on VIO, my bad. Hm, no more ideas from here, sorry. Hopefully someone else comes up with an idea.

PPPP · July 17, 2012, 3:58pm

That's alright zaxxon. Thanks for the quick responses

I'm just really frustrated...I don't have a lot of experience with AIX. I have a little more experience with Solaris and it's almost trivial to setup the network on that platform. I've been staring at the terminal for a few days but have no idea what the problem could be :wall:

---------- Post updated at 03:58 PM ---------- Previous update was at 03:55 PM ----------

Here's a question: how do I verify the interface I have on my LPAR is using the SEA? As in, what command(s) do I run on the LPAR to see if it's linked up correctly?

zaxxon · July 17, 2012, 5:11pm

I would suggest you compare via HMC the settings of the virtual ethernet adapter of one of your working LPARs with the non-working one.
Both have to be in a VLAN that is set on the SEA on the VIO. Your SEA can have more than 1 Virtual Lan ID. It needs to match at least one VLAN-ID between your LPAR's vAdapter and the VIO/SEA. It could also be, that all your client LPAR vAdapters are in the same VLAN which makes things a tad easier when comparing settings.
Maybe you have accidentally hooked a box for trunking or priority, whatever.
Have no HMC at hand, but here is a link, that shows exactly how things are set up:

http://www.ibm.com/developerworks/systems/library/es-pwr5-virtualvlan/

Foiled · July 17, 2012, 7:25pm

I would guess that your running ip tagging on your sea adapters. Which would mean when you setup your new lpar you would have to give it the correct vlan number for the vlan you addressed it in. On your lpar look at its properties then virtual adapters then your ethernet adapter...you should have a vlan number in there for "VLAN ID" and that vlan id number must be the correct one to match to one of the supported vlans ( which is a network ) on your SEA adapter on the vio. The vlan ID you use must be listed on the vio servers sea adapter properties and the vlan ID must correspond to the ip address you gave the server ( be in that subnet ).

polar · July 18, 2012, 2:04am

Hi, PPPP

make sure lpar AIX_SA vlan id same as AIX_CSS on HMC.
try to down detach ent0 on lpar AIX_CSS (if you are able to do so)

ifconfig en0 down detach

and lpar AIX_SA

ifconfig en1 down detach

start the tcp/ip for AIX_SA first and then for AIX_CSS.

zaxxon · July 18, 2012, 3:56am

I rephrased my former post from last night and added some info. Hopefully it is more helpful than before.

PPPP · July 18, 2012, 11:04am

Hey guys, thanks for all the responses

As I was reading through the post, going through the settings, I found a discrepancy between my non-working LPAR & the working one:

Under Hardware Information -> Virtual I/O Adapters -> Ethernet, on my non-working one (AIX_SA) it shows this:

On the working one (AIX_CSS), it shows this:

Notice the "Virtual Adapter" field for AIX_SA is empty. Could this be the reason why it's not working? How do I fix this?

Thanks!

---------- Post updated at 11:00 AM ---------- Previous update was at 09:19 AM ----------

That's exactly it. It is all on VLAN 1.

Thanks for the link! It goes through how to configure a SEA and how to hook things up properly. But how do I check if my current configuration is correct?

---------- Post updated at 11:04 AM ---------- Previous update was at 11:00 AM ----------

This could be another cause of my problem. The working LPAR is on a different subnet comparing to the non-working one. Forgive me for being an amateur but do they have to be on the same subnet? The VIO's SEA is on the same subnet as the working one. Do I need to create another SEA if my LPAR is on a different subnet?

PPPP · July 19, 2012, 4:54pm

Here's an update.

I tried to use DHCP and it worked! Then I reconfigured it back to use static IP and it stopped working again. After a while, it just started working I don't have time to figure out what I did to fix it at at this point but will check back later..

Thanks for all the help

zaxxon · July 19, 2012, 5:20pm

Sounds strange If you find out more, keep us updated please, thanks.