safe dragon
In MCC IP setup, we are seeing a difference in aggregate usage of approx. 15-20TB when comparing site A with Site B.
The aggregates are mirrored, metrocluster check is clean, plexes are online and in sync, vservers are healthy and there are same number of volumes on both sites. Is there anything other than storage efficiency and snapshots which might be causing this difference ?
upper fox
The aggregates on site A have nothing to do with the aggregates on site B. If you put more data on site A, then its aggregates will be fuller than the aggregates of site B
The mirroring in MetroCluster works on Plexes
this documentation and this TR for example explain how it works
safe dragon
Hi @upper fox , so if I understand this correctly, the output of 'aggr show' will only show the usage for the local plex (plex0). If Site A plex0 will be more utilized that Site B, then that's just how it is, it doesn't mean that the site B aggregate should also show the same usage ?
upper fox
almost. "aggr show" shows you the usage of the local aggregate only, yes. The local aggregate is composed of 2 mirrors, plex0 and plex1 (usually), and they are both exactly the same (bit-wise identical copies). plex1's disks are on the "remote" site (located in cluster B, so to say) but cluster B doesn't normally see them (unless in switchover). So the aggregate's "remote" mirror copy is completely invisible to cluster B even though the disks happen to be located "in" cluster B, if that makes sense