Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
xiemb
Explorer

The network was all down during the HA active/standby switchover

At the beginning, CP_B is the active and CP_A is the backup, after executing clusterXL_admin down,clusterXL_admin up on CP_B, the clients to access the server's business are all disconnected, and then executing clusterXL_admin down,clusterXL_admin up on CP_A, the clients to access the server's business are also disconnected. admin up on CP_A, the client to access the server is also disconnected. After looking at the messages, the status of HA is normal in both switchovers and there is also traffic to the firewall.
The topology diagram is as below:

0 Kudos
8 Replies
G_W_Albrecht
Legend Legend
Legend

 

That is possible for services that are not synchronized on the cluster. If connections are not set to keep data or rematch they will be disconnected. To check and synchronize a service, double click it => Advanced => Sync on cluster.

 

 

CCSP - CCSE / CCTE / CTPS / CCME / CCSM Elite / SMB Specialist
0 Kudos
xiemb
Explorer

Sync on cluster is on, I see traffic logs coming through the firewall, but noticed that new and concurrent connections are more than 3-4 times more than usual when switching over

0 Kudos
Lesley
Leader Leader
Leader

The increased amount of connections reflect the high load on both members.

Would further investigate the issue related to system load.

High load could impact cluster functionality. 

-------
If you like this post please give a thumbs up(kudo)! 🙂
0 Kudos
G_W_Albrecht
Legend Legend
Legend

You can decide for every service if and how to sync between cluster nodes.

CCSP - CCSE / CCTE / CTPS / CCME / CCSM Elite / SMB Specialist
0 Kudos
the_rock
Legend
Legend

I would definitely open TAC case to investigate this further. Just generate cpinfo files, as well as /var/log/messages*

Andy

0 Kudos
xiemb
Explorer

Thanks for your reply! I have exported the cpinf file the day after the failure. However, I see on the messages file that the firewall status is normal after switching active and standby twice.

0 Kudos
the_rock
Legend
Legend

Is clustering working right?

Please send below:

cphaprob state

cphaprob -a if

cphaprob roles

cphaprob -i list

cphaprob -l list

cphaprob syncstat

Andy

AmirArama
Employee
Employee

Maybe the network devices do not learn the ARP change of the VIP fast enough.

You can see if traffic still flowing on the GW after you run clusterXL_admin down

vmac might help if this is the case.

0 Kudos

Leaderboard

Epsum factorial non deposit quid pro quo hic escorol.

Upcoming Events

    CheckMates Events