'Valid HA node' vs 'NOT valid HA node'

[BCC]

I don't understand this question. How am I supposed to know which is the 'valid HA node' and which is the 'NOT valid HA node'? I don't see a "valid" field anywhere within the console.

[anton (staff)]

Valid HA node is the one who did not fall at all. Or fall last one (if both nodes are dead). In V5.7 we'll auto-track last processed write to handle this situation in automatic mode. For now you need to pick up the one yourself.

I don't understand this question. How am I supposed to know which is the 'valid HA node' and which is the 'NOT valid HA node'? I don't see a "valid" field anywhere within the console.

[BCC]

In a situation where both nodes are dead how do I tell which one failed last?

[anton (staff)]

1) It's not normal or common situation. At all...

2) Check "last written time" file attribute for image file (assuming you did not turn off clock synchronization for both cluster nodes...). Last written image is supposed to have the most recent data (but not guaranteed b/c of the obvious reasons).

3) In upcoming version we'll support "human factor" adding a) changes journalling b) triple node HA c) asynchronous mirroring for "last chance" recovery (de-duplication based).

In a situation where both nodes are dead how do I tell which one failed last?