Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
Christian_Koehl
Contributor

VSX: funny IP's in different subnets

Dear CheckMates,

I have some trouble with VSX VSLS in version R81.10 with Jumbo-30 (Mgmt is also on R81.10 with Jumbo-30).


For one of my eleven VS, I have lot's of error messages in /var/log/messages and the log file, saying "CLUS-1147-02-2: State chage: Active -> Standby...."

 

InterfaceMember-1Member-2
bond1.123192.168.196.33 / 28192.168.196.2 / 28
bond2.234192.168.196.17 / 28192.168.196.18 / 28
bond3.345192.168.196.1 / 28192.168.196.34 / 28

 

As we can see, the funny-ip's from bond1 and bond3 does not match the network ID of each other

Bond 2 is fine, as well as all other VS's.

 

Does someone has an idea to solve the problem?

 

Many thanks and best regars,

Christian

0 Kudos
3 Replies
Chris_Atkinson
Employee
Employee

Is this a new VS and are you in a position to remove and re-add the interfaces?

What was the order of operations that got you to this point?

Please engage TAC especially if the issue is reproducible.

0 Kudos
Christian_Koehl
Contributor

This is not a new VS. About 6-7 months ago we changed the appliance running the VSX Cluster from a 5900 units to a 26000 units and changed to configuration from bond1 and bond3.

This makes belive, we have the problem since that changes.

 

When running R80.40 with VSX HA everything was working. Now, we updated to R81.10 and now we running VSLS.

 

I had a TAC case in the past (when running R80.40), but they wanted to start debugging. Unfortunately, that systems are productive and the effected VS the "core router" of the customer. Therefore I  closed that case.

0 Kudos
Christian_Koehl
Contributor

Dear Folks,

last weekend I was able to have a close look into the problem.

It is possible to change the funny-ip's via GuiDBedit. The funny-ip's from the backup member were change. There are three positions for each funny-ip.

screenshot-1.png

 

screenshot-2.png

 

After saving the changes, GuiDBedit can be close. Now open the Smart Console and do a "Publish".

 

In the next step, I reset the VSX-cluster member with "reset_gw" via direct console connection.

The last step was a "vsx_util reconfigure" and a policy install.

 

I hope this will help, if someone else has this problem.

Best regards,

Christian