Resolving OSPF MTU problems with SROS

OSPF is a popular Interior Gateway Routing Protocol and in many instances it “just works” for a lot of situations, however care must be taken even in simple deployments. An issue that comes up from time to time is with regards to the maximum transmission unit (MTU). The network topology is a three router topology where I only have direct control a Nokia SROS based router.

OSPF MTU Test Topology

TL;DR – OSPF neighbor in ExchStart – you need to increase your MTU, OSPF neighbor in Exchange – you need to decrease your MTU. Keep reading to see how you can identify and resolve the MTU issues on Nokia Routers with SROS.

Below is the configuration of SR (the router under our administrative control):

configure
    system
        name "R1"
    exit
    sfm 1
        sfm-type m-sfm5-12
        no shutdown
    exit
    card 1
        card-type iom3-xp-b
        mda 1
            mda-type m10-1gb-xp-sfp
            no shutdown
        exit
        no shutdown
    exit
    port 1/1/1
        ethernet
            mode access
        exit
        no shutdown
    exit
    port 1/1/2
        ethernet
            mode access
        exit
        no shutdown
    exit
    router
        interface "system"
            address 1.1.1.1/32
            no shutdown
        exit
   exit
   service
      customer 1 create
          description "Default customer"
      exit
      ies 100 customer 1 create
          description "PEER1"
          interface "PEER1" create
              address 10.1.2.1/27
              sap 1/1/1 create
              exit
          exit
          no shutdown
      exit
      ies 200 customer 1 create
          description "PEER2"
          interface "PEER2" create
              address 10.1.3.1/27
              sap 1/1/2 create
              exit
          exit
          no shutdown
      exit
    exit
    router
        interface "system"
            address 1.1.1.1/32
            no shutdown
        exit
        ospf
            area 0.0.0.0
                interface "system"
                    no shutdown
                exit
           exit
           area 0.0.0.1
               interface "PEER1"
                   no shutdown
               exit
               interface "PEER2"
                   no shutdown
               exit
           exit
           no shutdown
       exit
   exit
exit all

Something to note is that the Peer routers are attached to an Internet Enhanced Service (IES) and not part of the OSPF Backbone Area – from a stored configuration perspective there is a distinction between core network and customer configurations but from a protocol perspective things are the same. IES Interfaces that are bound to Service Access Points (SAPs) which must be changed from the default mode of network – in this case we are using access, however hybrid is an option as well.

As this post is about resolving issues, obviously things are not working as straight forward as expected.

A:SR# show router route-table

===============================================================================
Route Table (Router: Base)
===============================================================================
Dest Prefix[Flags]                            Type    Proto     Age        Pref
    Next Hop[Interface Name]                                    Metric
-------------------------------------------------------------------------------
1.1.1.1/32                                    Local   Local     00h40m51s  0
   system                                                       0
10.1.2.0/27                                   Local   Local     00h33m32s  0
   PEER1                                                        0
10.1.3.0/27                                   Local   Local     00h34m02s  0
   PEER2                                                        0
-------------------------------------------------------------------------------
No. of Routes: 3
Flags: n = Number of times nexthop is repeated
B = BGP backup route available
L = LFA nexthop available
S = Sticky ECMP requested
===============================================================================

So nothing from OSPF is in the routing table, while its possible (but unlikely) that our peers aren’t advertising anything, an alternate explanation is that it could be a connectivity issue, so we’ll ping each peer router first

A:SR# ping 10.1.2.2 count 3
PING 10.1.2.2 56 data bytes
64 bytes from 10.1.2.2: icmp_seq=1 ttl=64 time=1.28ms.
64 bytes from 10.1.2.2: icmp_seq=2 ttl=64 time=1.15ms.
64 bytes from 10.1.2.2: icmp_seq=3 ttl=64 time=1.08ms.

---- 10.1.2.2 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.08ms, avg = 1.17ms, max = 1.28ms, stddev = 0.081ms
A:SR# ping 10.1.3.3 count 3
PING 10.1.3.3 56 data bytes
64 bytes from 10.1.3.3: icmp_seq=1 ttl=64 time=1.51ms.
64 bytes from 10.1.3.3: icmp_seq=2 ttl=64 time=1.31ms.
64 bytes from 10.1.3.3: icmp_seq=3 ttl=64 time=1.24ms.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.24ms, avg = 1.35ms, max = 1.51ms, stddev = 0.116ms

Okay so IP connectivity is established, lets check the OSPF interface state

A:SR# show router ospf interface

===============================================================================
Rtr Base OSPFv2 Instance 0 Interfaces
===============================================================================
If Name               Area Id         Designated Rtr  Bkup Desig Rtr  Adm  Oper
-------------------------------------------------------------------------------
system                0.0.0.0         1.1.1.1         0.0.0.0         Up   DR
PEER1                 0.0.0.1         1.1.1.1         100.100.100.100 Up   DR
PEER2                 0.0.0.1         1.1.1.1         0.0.0.0         Up   DR
-------------------------------------------------------------------------------
No. of OSPF Interfaces: 3
===============================================================================

From first glance PEER1 seems okay but PEER2 doesn’t have a BDR and since we are using the default ospf interface type (broadcast) we would expect that to see both the DR and BDR – lets get some more details

A:SR# show router ospf interface "PEER2" detail

===============================================================================
Rtr Base OSPFv2 Instance 0 Interface "PEER2" (detail)
===============================================================================
-------------------------------------------------------------------------------
Configuration
-------------------------------------------------------------------------------
IP Address       : 10.1.3.1
Area Id          : 0.0.0.1              Priority         : 1
Hello Intrvl     : 10 sec               Rtr Dead Intrvl  : 40 sec
Retrans Intrvl   : 5 sec                Poll Intrvl      : 120 sec
Cfg Metric       : 0                    Advert Subnet    : True
Transit Delay    : 1                    Cfg IF Type      : None
Passive          : False                Cfg MTU          : 0
LSA-filter-out   : None                 Adv Rtr Capab    : Yes
LFA              : Include              LFA NH Template  :
RIB-priority     : None
Auth Type        : None
-------------------------------------------------------------------------------
State
-------------------------------------------------------------------------------
Admin Status     : Enabled              Oper State       : Designated Rtr
Designated Rtr   : 1.1.1.1              Backup Desig Rtr : 0.0.0.0
IF Type          : Broadcast            Network Type     : Stub
Oper MTU         : 1500                 Last Enabled     : 06/02/2017 01:44:23
Oper Metric      : 100                  Bfd Enabled      : No
Te Metric        : 100                  Te State         : Down
Admin Groups     : None
Ldp Sync         : outOfService         Ldp Sync Wait    : Disabled
Ldp Timer State  : Disabled             Ldp Tm Left      : 0
-------------------------------------------------------------------------------
Statistics
-------------------------------------------------------------------------------
Nbr Count        : 0                    If Events        : 2
Tot Rx Packets   : 0                    Tot Tx Packets   : 76
Rx Hellos        : 0                    Tx Hellos        : 76
Rx DBDs          : 0                    Tx DBDs          : 0
Rx LSRs          : 0                    Tx LSRs          : 0
Rx LSUs          : 0                    Tx LSUs          : 0
Rx LS Acks       : 0                    Tx LS Acks       : 0
Retransmits      : 0                    Discards         : 78
Bad Networks     : 0                    Bad Virt Links   : 0
Bad Areas        : 78                   Bad Dest Addrs   : 0
Bad Auth Types   : 0                    Auth Failures    : 0
Bad Neighbors    : 0                    Bad Pkt Types    : 0
Bad Lengths      : 0                    Bad Hello Int.   : 0
Bad Dead Int.    : 0                    Bad Options      : 0
Bad Versions     : 0                    Bad Checksums    : 0
LSA Count        : 0                    LSA Checksum     : 0x0
===============================================================================

Okay, we can see that there are discards which align with the Bad Area Count – this means that PEER2 doesn’t believe it’s part of OSPF Area 1.

Log-id 99 is automatically configured on Nokia SROS devices to capture a number of event messages however it can get a bit overwhelming to find something specific. Fortunately there are ways to reduce the output by specifying the application (OSPF) and something that may be part of the log message itself we want to see (PEER2)

A:SR# show log log-id 99 application OSPF message PEER2

===============================================================================
Event Log 99
===============================================================================
Description : Default System Log
Memory Log contents  [size=500   next event=944  (wrapped)]

941 2017/06/02 02:10:15.55 UTC WARNING: OSPF #2043 Base VR:  1 OSPFv2 (0)
"LCL_RTR_ID 1.1.1.1: Conflicting configuration areaMismatch on interface PEER2 from 10.1.3.3 in hello"

So while we have identified a problem – OSPF Area MisMatch, we need to overcome it – remembering we cant configure PEER2 (the person that manages it is on a training course and cannot be contacted, while your project manager is wanting solutions, not problems..)

This is where using show and debug commands can help identify and resolve issues – SROS is quite powerful with its debugging tools and while they can be used in production, it is always best to attempt to narrow down what you are attempting to collect – firstly we need to create a debug log if one doesn’t already exist – for this example I’m just logging to a circular memory buffer but it could go to SNMP, syslog or a file if necessary.

A:SR# configure log log-id 10
*A:SR>config>log>log-id$ from debug-trace
*A:SR>config>log>log-id$ to memory
*A:SR>config>log>log-id$ no shutdown
*A:SR>config>log>log-id$ back
*A:SR>config>log# info
----------------------------------------------
log-id 10
    from debug-trace
    to memory
    no shutdown
exit
----------------------------------------------

Now to set up the debug – we know it’s from interface PEER2 and the log message kindly told us the packet type (in hello)..

*A:SR>config>log# /debug router ospf packet hello "PEER2"
*A:SR>config>log# show debug
debug
    router "Base"
        ospf
            packet hello "PEER2"
        exit
    exit
exit

Router Base is the global routing table of the router, the debug can reference other services e.g. a VPRN if necessary by changing the router – After a few seconds (OSPF hello packets will come every 10 seconds or so) we can look in log 10 to see what was received.

*A:SR>config>log# show log log-id 10

===============================================================================
Event Log 10
===============================================================================
Description : (Not Specified)
Memory Log contents  [size=100   next event=10  (not wrapped)]

9 2017/06/02 02:19:27.10 UTC MINOR: DEBUG #2001 Base OSPFv2
"OSPFv2: PKT

>> Outgoing OSPF packet on I/F PEER2 area 0.0.0.1
OSPF Version      : 2
Router Id         : 1.1.1.1
Area Id           : 0.0.0.1
Checksum          : ecb9
Auth Type         : Null
Auth Key          : 00 00 00 00 00 00 00 00
Packet Type       : HELLO
Packet Length     : 44 "

8 2017/06/02 02:19:26.55 UTC MINOR: DEBUG #2001 Base OSPFv2
"OSPFv2: PKT DROPPED
area mismatch"

7 2017/06/02 02:19:26.54 UTC MINOR: DEBUG #2001 Base OSPFv2
"OSPFv2: PKT

>> Incoming OSPF packet on I/F PEER2 area 0.0.0.2
OSPF Version      : 2
Router Id         : 200.200.200.200
Area Id           : 0.0.0.2
Checksum          : 5d27
Auth Type         : Null
Auth Key          : 00 00 00 00 00 00 00 00
Packet Type       : HELLO
Packet Length     : 44 "

SR1 is configured with PEER2 in Area 1 but it should be in Area 2, lets fix that

*A:SR>config>log# /configure router ospf
*A:SR>config>router>ospf# info
----------------------------------------------
area 0.0.0.0
    interface "system"
        no shutdown
    exit
exit
area 0.0.0.1
    interface "PEER1"
        no shutdown
    exit
    interface "PEER2"
        no shutdown
    exit
exit
no shutdown
----------------------------------------------
*A:SR>config>router>ospf# area 1 interface "PEER2" shutdown
*A:SR>config>router>ospf# area 1 no interface "PEER2"
*A:SR>config>router>ospf# area 2 interface "PEER2" no shutdown
*A:SR>config>router>ospf# info
----------------------------------------------
area 0.0.0.0
    interface "system"
        no shutdown
    exit
exit
area 0.0.0.1
    interface "PEER1"
        no shutdown
    exit
exit
area 0.0.0.2
    interface "PEER2"
        no shutdown
    exit
exit
no shutdown
----------------------------------------------

Now see if that fixes that problem.

*A:SR>config>router>ospf# show router ospf interface

===============================================================================
Rtr Base OSPFv2 Instance 0 Interfaces
===============================================================================
If Name               Area Id         Designated Rtr  Bkup Desig Rtr  Adm  Oper
-------------------------------------------------------------------------------
system                0.0.0.0         1.1.1.1         0.0.0.0         Up   DR
PEER1                 0.0.0.1         1.1.1.1         100.100.100.100 Up   DR
PEER2                 0.0.0.2         200.200.200.200 1.1.1.1         Up   BDR
-------------------------------------------------------------------------------
No. of OSPF Interfaces: 3
===============================================================================

Yes we can see both the DR and BDR for our OSPF peers but before we move on, we should stop the debug activity

*A:SR>config>router>ospf# /debug router no ospf
*A:SR>config>router>ospf# show debug
debug
exit

Now lets see if OSPF routing exchange is occurring.

*A:SR>config>router>ospf# show router route-table

===============================================================================
Route Table (Router: Base)
===============================================================================
Dest Prefix[Flags]                            Type    Proto     Age        Pref
Next Hop[Interface Name]                                    Metric
-------------------------------------------------------------------------------
1.1.1.1/32                                    Local   Local     01h15m46s  0
system                                                       0
10.1.2.0/27                                   Local   Local     01h08m27s  0
PEER1                                                        0
10.1.3.0/27                                   Local   Local     01h08m57s  0
PEER2                                                        0
-------------------------------------------------------------------------------
No. of Routes: 3
Flags: n = Number of times nexthop is repeated
B = BGP backup route available
L = LFA nexthop available
S = Sticky ECMP requested
===============================================================================

Well that isn’t fixed yet (which should be no surprise as this is about MTU issues) so lets move onto the next phase and examine the state of our OSPF neighbors

*A:SR>config>router>ospf# show router ospf neighbor

===============================================================================
Rtr Base OSPFv2 Instance 0 Neighbors
===============================================================================
Interface-Name                   Rtr Id          State      Pri  RetxQ   TTL
Area-Id
-------------------------------------------------------------------------------
PEER1                            100.100.100.100 ExchStart  1    0       34
0.0.0.1
PEER2                            200.200.200.200 Exchange   1    0       32
0.0.0.2
-------------------------------------------------------------------------------
No. of Neighbors: 2
===============================================================================

A router that is stuck in ExchStart or Exchange is a hallmark of OSPF MTU related problems.
Let’s start working on PEER1.

*A:SR>config>router>ospf# show router ospf neighbor "PEER1" detail

===============================================================================
Rtr Base OSPFv2 Instance 0 Neighbors for Interface "PEER1" (detail)
===============================================================================
-------------------------------------------------------------------------------
Neighbor : 10.1.2.2
-------------------------------------------------------------------------------
-------------------------------------------------------------------------------
Neighbor Rtr Id : 100.100.100.100  Interface: PEER1
-------------------------------------------------------------------------------
Neighbor IP Addr : 10.1.2.2
Local IF IP Addr : 10.1.2.1
Area Id          : 0.0.0.1
Designated Rtr   : 1.1.1.1              Backup Desig Rtr : 100.100.100.100
Neighbor State   : ExchStart            Priority         : 1
Retrans Q Length : 0                    Options          : - E - -  -  - - --
Events           : 1068                 Last Event Time  : 06/02/2017 02:59:34
Up Time          : 0d 01:11:00          Time Before Dead : 38 sec
GR Helper        : Not Helping          GR Helper Age    : 0 sec
GR Exit Reason   : None                 GR Restart Reason: Unknown (0)
Bad Nbr States   : 0                    LSA Inst fails   : 0
Bad Seq Nums     : 0                    Bad MTUs         : 1066
Bad Packets      : 0                    LSA not in LSDB  : 0
Option Mismatches: 0                    Nbr Duplicates   : 0
Num Restarts     : 0                    Last Restart at  : Never
===============================================================================

There are quite a few (1066) Bad MTUs being reported – While some vendors have an option to ignore the OSPF MTU, there are quite a number of MTU implications that can occur within the core when you consider various tunnel options that this is not provided.

Before we start to change things lets see if our trusty log 99 to see says anything about this:

*A:SR>config>router>ospf# show log log-id 99 application OSPF message PEER1

===============================================================================
Event Log 99
===============================================================================
Description : Default System Log
Memory Log contents  [size=500   next event=1473  (wrapped)]

1472 2017/06/02 02:40:17.67 UTC WARNING: OSPF #2043 Base VR:  1 OSPFv2 (0)
"LCL_RTR_ID 1.1.1.1: Conflicting configuration mtuMismatch on interface PEER1 from 10.1.2.2 in dbDescript"

We can use another debug to determine what the actual MTU should be (as before with the area mismatch, log 99 gave us a hint as to the packet type we should be investigating):

*A:SR>config>router>ospf# /debug router ospf packet dbdescr ingress "PEER1"

Clear the log and see what we are receiving:

*A:SR>config>router>ospf# /clear log 10
*A:SR>config>router>ospf# /show log log-id 10

===============================================================================
Event Log 10
===============================================================================
Description : (Not Specified)
Memory Log contents  [size=100   next event=3  (not wrapped)]

2 2017/06/02 02:55:17.67 UTC MINOR: DEBUG #2001 Base OSPFv2
"OSPFv2: PKT DROPPED
MTU mismatch"

1 2017/06/02 02:55:17.67 UTC MINOR: DEBUG #2001 Base OSPFv2
"OSPFv2: PKT

>> Incoming OSPF packet on I/F PEER1 area 0.0.0.1
OSPF Version      : 2
Router Id         : 100.100.100.100
Area Id           : 0.0.0.1
Checksum          : e35a
Auth Type         : Null
Auth Key          : 00 00 00 00 00 00 00 00
Packet Type       : DB_DESC
Packet Length     : 32

Interface MTU     : 1504
Options           : 000042
Flags             : 7   INIT MORE MAST
Sequence Num      : 2514
"

Okay, so PEER1 requires an MTU of 1504, lets modify that within the OSPF configuration:

*A:SR>config>router>ospf# info
----------------------------------------------
area 0.0.0.0
    interface "system"
        no shutdown
    exit
exit
area 0.0.0.1
    interface "PEER1"
        no shutdown
    exit
exit
area 0.0.0.2
    interface "PEER2"
        no shutdown
    exit
exit
no shutdown
----------------------------------------------
*A:SR>config>router>ospf# area 1 interface "PEER1" mtu 1504

When applying a configuration it is good to verify things are working as expected:

*A:SR>config>router>ospf# show router ospf interface "PEER1" detail | match MTU
Passive          : False                Cfg MTU          : 1504
Oper MTU         : 1500                 Last Enabled     : 06/02/2017 01:44:23

Although we configured the MTU to be 1504, the Operational MTU is 1500 (This is because the IP MTU is 1500 so OSPF cant be given a larger MTU on this interface)

*A:SR>config>router>ospf# show router interface "PEER1" detail | match MTU
IP MTU           : (default)
IP Oper MTU      : 1500

When Ethernet Ports are configured as mode access and left at the default encapsulation (null) the Ethernet port MTU is 1514 bytes (to support a 1500 byte IP MTU and 14 bytes of Ethernet Header – FCS is not included in MTU calculations)

*A:SR>config>router>ospf# show port 1/1/1 | match MTU
Physical Link      : Yes                        MTU              : 1514

To get a 1504 byte IP MTU, we can just add 4 bytes to the Port Ethernet MTU

*A:SR>config>router>ospf# /configure port 1/1/1 ethernet mtu 1518
*A:SR>config>router>ospf# show router interface "PEER1" detail | match MTU
IP MTU           : (default)
IP Oper MTU      : 1504
*A:SR>config>router>ospf# show router ospf interface "PEER1" detail | match MTU
Passive          : False                Cfg MTU          : 1504
Oper MTU         : 1504                 Last Enabled     : 06/02/2017 01:44:23

This should mean that the OSPF neighbor will now perform the database exchange and enter the Full state.

*A:SR>config>router>ospf# show router ospf neighbor "PEER1"

===============================================================================
Rtr Base OSPFv2 Instance 0 Neighbors for Interface "PEER1"
===============================================================================
Interface-Name                   Rtr Id          State      Pri  RetxQ   TTL
Area-Id
-------------------------------------------------------------------------------
PEER1                            100.100.100.100 Full       1    0       34
0.0.0.1
-------------------------------------------------------------------------------
No. of Neighbors: 1
===============================================================================
*A:SR>config>router>ospf# show router route-table

===============================================================================
Route Table (Router: Base)
===============================================================================
Dest Prefix[Flags]                            Type    Proto     Age        Pref
Next Hop[Interface Name]                                    Metric
-------------------------------------------------------------------------------
1.1.1.1/32                                    Local   Local     02h10m36s  0
system                                                       0
10.1.2.0/27                                   Local   Local     02h03m16s  0
PEER1                                                        0
10.1.3.0/27                                   Local   Local     02h03m46s  0
PEER2                                                        0
100.100.100.100/32                            Remote  OSPF      00h03m18s  10
10.1.2.2                                                     100
-------------------------------------------------------------------------------
No. of Routes: 4
Flags: n = Number of times nexthop is repeated
B = BGP backup route available
L = LFA nexthop available
S = Sticky ECMP requested
===============================================================================

The OSPF issue with PEER1 appears to have been resolved so back to PEER2.

*A:SR>config>router>ospf# show router ospf neighbor "PEER2" detail

===============================================================================
Rtr Base OSPFv2 Instance 0 Neighbors for Interface "PEER2" (detail)
===============================================================================
-------------------------------------------------------------------------------
Neighbor : 10.1.3.3
-------------------------------------------------------------------------------
-------------------------------------------------------------------------------
Neighbor Rtr Id : 200.200.200.200  Interface: PEER2
-------------------------------------------------------------------------------
Neighbor IP Addr : 10.1.3.3
Local IF IP Addr : 10.1.3.1
Area Id          : 0.0.0.2
Designated Rtr   : 200.200.200.200      Backup Desig Rtr : 1.1.1.1
Neighbor State   : Exchange             Priority         : 1
Retrans Q Length : 3                    Options          : - E - -  -  - O --
Events           : 3                    Last Event Time  : 06/02/2017 02:22:36
Up Time          : 0d 01:01:10          Time Before Dead : 37 sec
GR Helper        : Not Helping          GR Helper Age    : 0 sec
GR Exit Reason   : None                 GR Restart Reason: Unknown (0)
Bad Nbr States   : 0                    LSA Inst fails   : 0
Bad Seq Nums     : 0                    Bad MTUs         : 0
Bad Packets      : 0                    LSA not in LSDB  : 0
Option Mismatches: 0                    Nbr Duplicates   : 917
Num Restarts     : 0                    Last Restart at  : Never
===============================================================================

There are no Bad MTUs being reported here, all we can see is that we are forever in Exchange state – lets check log 99 to see if anything at all related to PEER2 is present

*A:SR>config>router>ospf# show log log-id 99 message PEER2

===============================================================================
Event Log 99
===============================================================================
Description : Default System Log
Memory Log contents  [size=500   next event=2033  (wrapped)]

There is nothing present (the older events have wrapped around since we are only keeping the last 500 events)
What I have found is when OSPF neighbors are stuck in ExchStart, your router is the one with the MTU too small but while the router that is stuck in Exchange is the one with the MTU that is too big for its peer.
To work out what the smaller MTU should be, we’ll send ping packets of various lengths to work out what is the biggest unfragmented packet that can be sent to PEER2. Note: when we send a ping and specify the size, we are actually calling out what the ICMP payload size should be, so we need to ensure for IPv4 we consider the 20 byte IP header and 8 byte ICMP header – so an IP interface with an IP-MTU of 1500 would work for a ping with a payload size of 1472 but would fail at 1473.
We can test this concept on a known quantity (PEER1 which has an IP MTU of 1504) we should be able to get a ping payload of 1476 through okay but 1477 should fail – make sure we set the DF bit!

*A:SR>config>router>ospf# ping 10.1.2.2 size 1476 do-not-fragment count 3
PING 10.1.2.2 1476 data bytes
1484 bytes from 10.1.2.2: icmp_seq=1 ttl=64 time=1.34ms.
1484 bytes from 10.1.2.2: icmp_seq=2 ttl=64 time=1.27ms.
1484 bytes from 10.1.2.2: icmp_seq=3 ttl=64 time=1.16ms.

---- 10.1.2.2 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.16ms, avg = 1.26ms, max = 1.34ms, stddev = 0.072ms
*A:SR>config>router>ospf# ping 10.1.2.2 size 1477 do-not-fragment count 3
PING 10.1.2.2 1477 data bytes

---- 10.1.2.2 PING Statistics ----
3 packets transmitted, 3 packets bounced, 0 packets received, 100% packet loss

This works as expected, so the concept appears sound.

*A:SR>config>router>ospf# show router interface "PEER2" detail | match MTU
IP MTU           : (default)
IP Oper MTU      : 1500

We know that we have a ceiling of 1500 and we know the MTU must be lower than this. But just to be certain, we’ll try based on a 1500 byte IP packet anyway

*A:SR>config>router>ospf# ping 10.1.3.3 size 1472 do-not-fragment count 3
PING 10.1.3.3 1472 data bytes
Request timed out. icmp_seq=1.
Request timed out. icmp_seq=2.
Request timed out. icmp_seq=3.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 0 packets received, 100% packet loss

Unsurprising, the Peer MTU is less than 1500 bytes, lets try a slightly smaller payload

*A:SR>config>router>ospf# ping 10.1.3.3 size 1462 do-not-fragment count 3
PING 10.1.3.3 1462 data bytes
1470 bytes from 10.1.3.3: icmp_seq=1 ttl=64 time=1.44ms.
1470 bytes from 10.1.3.3: icmp_seq=2 ttl=64 time=1.36ms.
1470 bytes from 10.1.3.3: icmp_seq=3 ttl=64 time=1.34ms.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.34ms, avg = 1.38ms, max = 1.44ms, stddev = 0.042ms

Okay, time to divide and conquer to determine the largest payload that gets through

*A:SR>config>router>ospf# ping 10.1.3.3 size 1467 do-not-fragment count 3
PING 10.1.3.3 1467 data bytes
1475 bytes from 10.1.3.3: icmp_seq=1 ttl=64 time=1.24ms.
1475 bytes from 10.1.3.3: icmp_seq=2 ttl=64 time=1.30ms.
1475 bytes from 10.1.3.3: icmp_seq=3 ttl=64 time=2.16ms.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.24ms, avg = 1.57ms, max = 2.16ms, stddev = 0.417ms
*A:SR>config>router>ospf# ping 10.1.3.3 size 1469 do-not-fragment count 3
PING 10.1.3.3 1469 data bytes
Request timed out. icmp_seq=1.
Request timed out. icmp_seq=2.
Request timed out. icmp_seq=3.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 0 packets received, 100% packet loss
*A:SR>config>router>ospf# ping 10.1.3.3 size 1468 do-not-fragment count 3
PING 10.1.3.3 1468 data bytes
1476 bytes from 10.1.3.3: icmp_seq=1 ttl=64 time=1.19ms.
1476 bytes from 10.1.3.3: icmp_seq=2 ttl=64 time=1.30ms.
1476 bytes from 10.1.3.3: icmp_seq=3 ttl=64 time=1.26ms.

---- 10.1.3.3 PING Statistics ----
3 packets transmitted, 3 packets received, 0.00% packet loss
round-trip min = 1.19ms, avg = 1.25ms, max = 1.30ms, stddev = 0.043ms

An ICMP payload of 1468 fits within an IP packet with a size of 1496 (add 8 Bytes for ICMP Header and 20 Bytes for IP Header) – adjust the OSPF MTU to 1496 and see if that results in getting a full adjacency.

*A:SR>config>router>ospf# area 2 interface "PEER2" mtu 1496
*A:SR>config>router>ospf# show router ospf interface "PEER2" detail | match MTU
Passive          : False                Cfg MTU          : 1496
Oper MTU         : 1496                 Last Enabled     : 06/02/2017 02:22:36
*A:SR>config>router>ospf# show router ospf neighbor "PEER2"

===============================================================================
Rtr Base OSPFv2 Instance 0 Neighbors for Interface "PEER2"
===============================================================================
Interface-Name                   Rtr Id          State      Pri  RetxQ   TTL
Area-Id
-------------------------------------------------------------------------------
PEER2                            200.200.200.200 Full       1    0       35
0.0.0.2
-------------------------------------------------------------------------------
No. of Neighbors: 1
===============================================================================

The adjacency is up – lets see what routes we have learnt

*A:SR>config>router>ospf# show router route-table

===============================================================================
Route Table (Router: Base)
===============================================================================
Dest Prefix[Flags]                            Type    Proto     Age        Pref
Next Hop[Interface Name]                                    Metric
-------------------------------------------------------------------------------
1.1.1.1/32                                    Local   Local     02h40m46s  0
system                                                       0
10.1.2.0/27                                   Local   Local     02h33m27s  0
PEER1                                                        0
10.1.3.0/27                                   Local   Local     02h33m56s  0
PEER2                                                        0
100.100.100.100/32                            Remote  OSPF      00h33m29s  10
10.1.2.2                                                     100
200.200.200.200/32                            Remote  OSPF      00h01m27s  10
10.1.3.3                                                     100
-------------------------------------------------------------------------------
No. of Routes: 5
Flags: n = Number of times nexthop is repeated
B = BGP backup route available
L = LFA nexthop available
S = Sticky ECMP requested
===============================================================================

We now have learnt routes from PEER1 and PEER2, time for a quick dataplane verification:

*A:SR>config>router>ospf# ping 100.100.100.100 source 1.1.1.1 count 1
PING 100.100.100.100 56 data bytes
64 bytes from 100.100.100.100: icmp_seq=1 ttl=64 time=1.38ms.

---- 100.100.100.100 PING Statistics ----
1 packet transmitted, 1 packet received, 0.00% packet loss
round-trip min = 1.38ms, avg = 1.38ms, max = 1.38ms, stddev = 0.000ms
*A:SR>config>router>ospf# ping 200.200.200.200 source 1.1.1.1 count 1
PING 200.200.200.200 56 data bytes
64 bytes from 200.200.200.200: icmp_seq=1 ttl=64 time=1.14ms.

---- 200.200.200.200 PING Statistics ----
1 packet transmitted, 1 packet received, 0.00% packet loss
round-trip min = 1.14ms, avg = 1.14ms, max = 1.14ms, stddev = 0.000ms

We now have successful routing exchange and data plane reachability.