Solarwinds 2022.3 instability due to HA failure

After I have upgraded my solarwinds version to 2022.3 , the platform has become very unstable and experience a lot monitoring downtime.

the problem is the primary server failover to the secondary, but when the standby server hand over the service back to the primary preferred server , services are disabled and not running and could not be started . Even if the manually started it will disable them.

tried working several Support personnel changing manually the DB tables to make server standby and active mode , modifying registry etc . after they restored the services it will reoccur again. due to this I am considering downgrading

why would the service has manual start up on primary and disabled startup on standby server on services ?

is there an option to wait or tolerate communication failure , due to delay etc before it fails beyond 60 seconds ?

when it fails the web console transitioning to the secondary server very long time (>10 min). is there any way of improving this ?

Thanks

Find more posts tagged with

orion

Accepted answers

All comments

riggs.clough

Hello,

Can you please refer me to your support ticket #? We'd definitely like to look into the diagnostics on the HA pair to see what is preventing a smooth HA failover.

Thanks

tony.johnson

Hello, Could you share the support case reference with me and I will escalate this to our engineering team for further investigation.

mestasew

Thanks Tony -

send me a private message so that I can share you a support case numbers that I have already uploaded diagnostics.

Any ways discussing the future on the HA - why would the standby server has a disabled start up setting on the server ? Rather it make sense a server be set in automatic startup and when HA determine a peer down it will change the stopped services to running state.

mestasew

Hello Riggs- please reach me via private message for sharing.

guhmorello

Hello,

I'm having the same problem on 2 clients I support.

I haven't opened a call for both of them yet, but I'm sure I will.

@mestasew if your case is successful, please post the answer here.

thanks,

Best Regards

guhmorello

Hello,

Yesterday while checking this problem I found that the two servers, MainPoll and MainPollStandby were getting the same entries in the database, HA_PoolMembers tables.

When MainPoll started the FailOver process, it changed the database entry to MainPollStandby and that was how the two were.

Searching I found the following RabbitMQ procedure and item 6 solved the problem:

6 - Troubleshoot the inconsistent_cluster error
support.solarwinds.com/.../Rabbit-MQ

After performing the above procedure, access to the Web environment was restored and for security purposes I left the HA service disabled on the standby server.

thank you,

mestasew

sorry for the delayed comment. HA availability was established after breaking the ha by truncating the HA table, modified DB tables and even registry to assign values of the server roles as primary or standby. Finally reestablishing HA.

still we have some issue displaying web console when it failover to standby and the failover process is slow. Standby server took over services in reasonable time but access to web console is not working causing standard users deteriorated web console access experience.

guhmorello

Hello,

Mestasew, you have any solution for this Topic?

thank you and have a nice week!