Potentially unexpected redundancy behavior?

Ignition
,
1

Using 8.1.25.

I have two servers in redundancy that host a vision application. Had unexpected behavior last time they had to patch the primary server so just wanted to clarify a few things.

When I manually failover, it lasts until I fail things back over and I see on the status page the status change from Backup to Master or vice versa.

  1. However when server 1 was given a hard reset to run its updates without being failed over first - on server 2 the status page never changed so it didn’t seem clear that server 2 took over although vision users did increase and the instance of the applicaiton I had seemed to fail over ok. Normal to not see the status change in this event?

  2. Also, when server 1 came back online, it resumed status as the master which I was not expecting - is this expected behavior?

  3. A general question to people who do use it, does the redundancy not work nicely with the master getting a hard shut off, I know its better to gracefully failover when you can but has anyone experienced issues from unexpected failovers?

2
  1. This could be related to the Backup Node Settings (if Ping Max Missed is a high value and/or the Ping Timeout is high).
  2. This depends on the Master Node Settings (Automatic or Manual).
  3. I specifically recall issues between 8.1.18-8.1.24 with a redundant install I was working on. 8.1.25 introduced a proper fixed that allowed redundant pairs to be on different versions for the sake of upgrading. Since, then I’ve not any issues like that.

8.1.25 snippet:

image
image1036×258 20.9 KB

1 Like