9100 with compression?
Yeah - I've seen that configuration deployed. I've seen it seem OK in testing then catastrophically fail four days into production use, with low IOPS (1-2k IOPS) and relatively small data volumes (6TB total), where the underlying storage was corrupted and all LUNs were lost. Haven't seen the port-mortem yet but the rumour was the compression code did something funny. Might be worth pressing your IBM rep if you're thinking of it.
Biting the hand that feeds IT
