VSLS - Documentation on 'Active Up' and 'Higher Up' modes.

[JasonMcAllister]

Hi All,

I'm trying to find some documentation which describes the differences and use cases for the option to toggle between Active Up and Higher Up VSLS modes in vsx_util.

I've checked through VSLS documentation in the R80.30 VSX admin guide but it seems not to be even mentioned.

Does anyone have details on what the use cases between each would be and what they actually do/differ?

Many thanks,

Jason

[PhoneBoy · Accepted solution]

It controls the behavior of what happens when the primary active member for a VS becomes active after a failure.
In regular ClusterXL (non-VSX), the primary member becomes the primary for the cluster once again.
In VSX, the default is to not fail back to the primary when it comes back up.
This "HigherUp" option changes the behavior in VSLS to be like regular ClusterXL.

[JasonMcAllister]

Thank you @PhoneBoy for your response.

In summary:

Active Up - The VS that has failed over will remain in place and not return to the original recovered member.

Higher Up - The VS, like regular ClusterXL, will return.

My experience in VSLS, however, is that the load sharing weights and preferred orders of a VS will move to be active on the recovered member if it was - Which is what I expect as it is Load Sharing and not just availability.

Is this behaviour change more applicable to VSHA?

[PhoneBoy]

That may very well be the case.
I found a very brief explanation of this in a couple of TAC cases and was trying to extrapolate it.