I just got back from a great LeftHand Networks training course. I've been a big fan of the VSA running on my Lab Network for a while. Here are my rough notes from the class.
Each LeftHand Storage Module has 2 NICs.
Snapshots are Read Write volumes. This is by default and different from most vendor snapshots which are Read Only.
Types of NIC Aggregation supported by LeftHand SANs:
- Active-Passive or Active-Backup
- Link Agregation – 802.11AD
- Adaptive Load Balancing: 1 GB in and 2 GB outbound. (Most commonly used)
LeftHand Tools: Health Check Utility – Monitors system Logs and optional emails them to support and client on a scheduled basis. Helpful when you call for support, they have your logs already (or at least a version of them)
For High Availability: You should run 3 managers or 5 Managers (the Most Managers recommended). Interestingly enough, 2 VSAs are not enough for a COMPLETELY automated fail over. Your data will be safe but your LUN will be offline until you fire up the Virtual Manager on the surviving node and bring the LUN back online.
Virtual Manager is used to regain quorum for a replicated LUN.
DiskPar and DiskPart are used for aligning the partition table offset correctly.
create partition primary align =64
This tidbit was extremely useful. It was said that misaligned disks can suffer from a 2 – 20% performance hit depending on usage.
Disk alignment will be automatic with Windows 2008.
Some of the items from the Certification Test:
- Soft shutdown is an aspect of the management group.
- Delete management group when you are COMPLETELY redoing your storage.
- LUNs and Volumes are the Same.
- Initiator Node Name is always retrieved from the initiator.
- Remote bandwidth is what the managers will dedicate to remote copy.
- Replication is Snapshot to snapshot only
- Oracle control files must be on Direct Attached…
- Managers keep track of everything…