Ospf 3.10 BUG: Database Description packet has different mas

RouterOS Forwarding Protocols
1

We have one ptp wireless link with two ROS 3.10 on both end. First(ospf id 10.11.0.51) is closer to the network center. Second (ospf id 10.11.0.69) is more distant from center. There are only these two routers on that link. Sometimes the route to second router disappears from all routers in the area. Only solution is to restart ospf on the second router. This is a log from the first router

20:08:16 route,ospf,info OSPF neighbor 10.11.0.69: state change from Exchange to
 2-Way 
20:08:57 route,ospf,info OSPF neighbor 10.11.0.69: state change from Exchange to
 2-Way 
20:09:38 route,ospf,info OSPF neighbor 10.11.0.69: state change from Exchange to
 2-Way 
20:10:19 route,ospf,info OSPF neighbor 10.11.0.69: state change from Exchange to
 2-Way 
20:11:01 route,ospf,info OSPF neighbor 10.11.0.69: state change from Exchange to
 2-Way

And this is a log from second router.

20:12:26 route,ospf,info Database Description packet has different master status flag 
20:12:26 route,ospf,info     new master flag=false 
20:12:26 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:13:08 route,ospf,info Database Description packet has different master status flag 
20:13:08 route,ospf,info     new master flag=false 
20:13:08 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:13:49 route,ospf,info Database Description packet has different master status flag 
20:13:49 route,ospf,info     new master flag=false 
20:13:49 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:14:30 route,ospf,info Database Description packet has different master status flag 
20:14:30 route,ospf,info     new master flag=false 
20:14:30 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:15:11 route,ospf,info Database Description packet has different master status flag 
20:15:11 route,ospf,info     new master flag=false 
20:15:11 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:15:52 route,ospf,info Database Description packet has different master status flag 
20:15:52 route,ospf,info     new master flag=false 
20:15:52 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way 
20:16:33 route,ospf,info Database Description packet has different master status flag 
20:16:33 route,ospf,info     new master flag=false 
20:16:33 route,ospf,info OSPF neighbor 10.11.0.51: state change from Full to 2-Way

Any suggestions?

Thank you.

ospf NSSA extremely unstable while stubs are great
2

upgrade routeros to latest version.
There have been close to a dozen fixes to ospf since 3.10.

http://www.mikrotik.com/download/CHANGELOG_3

3

I updated routers to 3.24 but there are several links where the same error still occurs.

20:12:26 route,ospf,info Database Description packet has different master status flag
20:12:26 route,ospf,info new master flag=false

I tried totally change 6 links to ospf from rip and i have had problem with 3 of them so it seems strange. I only add network address to the backbone and change ospf router Id. Only solution is to stop ospf on both end of the link and then start it again. This solution doesn’t last long.

4

Hello,
i have the same error with ROS 3.30 on both Routers.

Any help?

Greetings Manuel Ritter

5

You guys won’t get any help until you upgrade to the latest RouterOS release.

6

Same error on RouterOS 4.5

[EDIT]
I had this error 2 times after the upgrade of both routers, but now it seems to be gone.
[/EDIT]

7

Any updates on this?
I’ve tried changing from broadcast to NBMA, upgrading to 4.5.. no luck.. :frowning:

8

Only seems to be one particular PTP link as well..

9

Was there ever anything found on this? This problem just popped up on a link that’s been good for quite some time. both running RouterOS V4.13 (Power Router 732s). Just wondering before I head out there tomorrow to change ports to see if that helps.

10

Actually, this ended up being from loss of comunication between the routers due to one of the routers constantly rebooting due to cold weather. That is why “state changes” only counted on one side.

11

I’ve the same problem with router rb1100 and rb450g and rb435g with current os 5.14… would mikrotik fix it. 'Cause we need it so much. Would you do me a favor for this

12

The problem, BUG instead, persists on version 5.17, this is very annoying, if someone knows an easy workaround would be nice !
i’ve a mesh network topology, it consists of over 50 wireless nodes connected together, some link is p2p someone p2mp, ospf network-type is broadcast. actually every day at least one node lose routing to an horizon and in most cases i observed that node was p2p (a simple wireless AP-Station not wds)

13

Because of the MUlticast problems on wireless,
One of Solutions is to use NBMA
http://wiki.mikrotik.com/wiki/Manual:OSPF-examples#NBMA_networks

Another solution use wds or new station bridge mode and multicast helper on ap (works only between mikrotik devices).
More info in the manual
http://wiki.mikrotik.com/wiki/Manual:Multicast_detailed_example#Multicast_and_Wireless

14

WDS solve nothing

15

This happen almost all the time, if link between 2 roters drop. Its not just wireless. Fiber, copper, wireless, etc etc, if a link drop, and 2 routers is stil up, when link is back online, it seems like ospf dont want to connect agane. Also the same in nbma mode.

This is a really big bug, That mikrotik dont want to beleve, but it still happen to often.

also, se :

http://forum.mikrotik.com/t/database-description-packet-has-different-master-status-flag/30658/1

16

+1 more

17

same problem v5.18

BTW next bug in OSPFv3

20:03:07 route,ospf,error Discarding packet: locally originated 
20:03:07 route,ospf,error     src address=fe80::225:90ff:fe27:a6d
18

Hi,

I have this problem in 5.18 - One box looses two ospf neighbours at the same time and the other neighbours sees that there is a different master flag.

Since this happens at the same time with two directly connected neighbours its little wierd. Routers are using a unique (checked too) ip on a loopback as router id.

I have changed the router (was powerboard 732 - to RB1100 - to RB1100X2).

I dont see any issues on the ethernet.

Any clues or is this a bug?

/M

19

++1

20

+++1