XSAN // Multiple switches for redundancy
hi guys,
we have a xsan setup running for some time successfully now and recently added a new, second fc switch to have a failover/redundancy in case one fails.
unfortunately this is not working properly and also speed/throughput is not as good as before with the second switch.
unfortunately I'm a newbee so some help would be very appreciated 🙂
2x xserve mdc (each with apple dual fc hba 2x4GB)
5x xserve clients (each with atto dual fc hba 2x8GB)
2x qlogic 5800
1x promise vtrak (2 ctrl.)
1x promise vtrak jbod (attached to above)
we have created 5 luns on the storage
data-1 (xsan lun1, xsan lun2, each 7tb)
data-2 (xsan lun3, xsan lun4, each 7tb)
metadata-journal (xsan lun5, 2 tb)
lun1-4 are raids 1R, 7 hdd x 2tb
lun5 (metadata) is a raid 1 (mirror), 2 hdd x 2tb
+2 global spares
we figured out that this is the most performant setup for our requirements.
with everything connected to a single switch we reach on the clients with the 8GB hbas up to 1.5 GB/sec.
now I've added the second switch, connected each HBA to both switches and also connected the promise raid to both switches (2x switch 1, 2x switch 2), but the data rates seem to be only 50%, and also when I pull the power plug from one switch, the xsan goes offline.
is it a problem with mapping? Or have I wired it wrong?
we have a xsan setup running for some time successfully now and recently added a new, second fc switch to have a failover/redundancy in case one fails.
unfortunately this is not working properly and also speed/throughput is not as good as before with the second switch.
unfortunately I'm a newbee so some help would be very appreciated 🙂
2x xserve mdc (each with apple dual fc hba 2x4GB)
5x xserve clients (each with atto dual fc hba 2x8GB)
2x qlogic 5800
1x promise vtrak (2 ctrl.)
1x promise vtrak jbod (attached to above)
we have created 5 luns on the storage
data-1 (xsan lun1, xsan lun2, each 7tb)
data-2 (xsan lun3, xsan lun4, each 7tb)
metadata-journal (xsan lun5, 2 tb)
lun1-4 are raids 1R, 7 hdd x 2tb
lun5 (metadata) is a raid 1 (mirror), 2 hdd x 2tb
+2 global spares
we figured out that this is the most performant setup for our requirements.
with everything connected to a single switch we reach on the clients with the 8GB hbas up to 1.5 GB/sec.
now I've added the second switch, connected each HBA to both switches and also connected the promise raid to both switches (2x switch 1, 2x switch 2), but the data rates seem to be only 50%, and also when I pull the power plug from one switch, the xsan goes offline.
is it a problem with mapping? Or have I wired it wrong?
