Improve Storage Efficiency and Management with VMware vSphere 4 Redux
Tuesday, June 23, 2009 at 2:06PMContinuing on, here is a quick synopsis of the following webcast, http://www.vmware.com/a/webcasts/details/261, title above.
- So what’s new you ask? Storage efficiency as it relates to the following…
- New iSCSI software initiator
- PVSCSI adapter
- VMDirectPath IO
- Virtual disk thin provisioning
- Improved management
- vStorage Thin Provisioning
- Consumes only what is used (over commitment)
- Fully supported on block storage
- VM sees full logical disk size at all times
- full reporting and alerting on allocation and consumption
- improves storage utilization
- eliminates over provisioning
- Via enhanced Storage vMotion gives you the ability to migrate thick to thin disks
- Potential negative? Metadata updates needed more frequently with thin disks
- Fragmentation a problem? incremental size increase is based on block size of VMFS volume. Lower block sizes less of an issue.
- TP Option available at: VM creation, clone to a template, clone a VM, migrate a VM (Storage Motion)
- Reporting and alerting are key, Rate of disk growth (writes specifically) important to note for thin disk. Ultimately this can affect performance.
- Eager zeroed thick disk required for FT, thick lazy zeroed default for VM creation
-
Additional Datastore Management built around the following:
-
Added Datastore views, you can now manage Datastores independent of the hosts.
-
Much need usage reports on a per DS level
-
The ability to set alarms/alerts on capacity, and group DS’s as foldered objects. Not only that you can set permissions on who can allocate to certain DS’s. Bitchin’..
-
-
Dynamic expansion-VMFS volume grow
-
Current options to alleviate oversubscription
-
increase the Datastore size or VMFS Volume (add extend, span, grow the VMFS volume)
-
Storage vMotion
-
Cold migrate of VM to another Datastore
-
Delete old and unused VM’s from Datastore
-
-
VMFS volume grow allows you to expand an extend so that it fills the available adjacent capacity there fore improving VM availability
-
Can grow as many times as needed up to max volume size of the VMFS volume
-
Must grow LUN backing for the VMFS Datastore first. So LUN expansion and then the VMFS volume grow
-
Volume LUN size is still maxed at 2TB. This stems from SCSI 2 compliancy with in the vmkernel.
-
Storage VMotion now supports NFS/iSCSI and FC and fully GUI integration. local experimental
-
-
-
Enhanced Storage vMotion
- Storage VMotion now supports NFS/iSCSI and FC and is fully GUI integration. Local disk migrations are experimental.
- Changes within the mechanisms behind this function
- Snapshot in 3.x to do sVMotion, in vSphere however, enhanced Storage VMotion flows as follows.
- Copy VM home to new location
- Start changed block tracking
- Pre-copy disk to destination (multiple iterations)
- Copy all remaining disk blocks Fast suspend/resume
- VM to start running on new home and disks
- Delete original VM and disks
- Furthermore in vSphere, there is support for moving VMDK’s from thick to thin formats and migrating RDM’s to VMDK’s
- Software iSCSI initiator
- No longer requires Service Console connection for initiating communication with iSCSI target.
- Additional performance tuning parameters under advanced within the initiator properties
- CPU cost improvement---read +10 to 25% improvement, write 20 to 50% improvement
- Jumbo frames and 10G support for the TCP/IP stack
- vStorage API for multipathing
- Pluggable storage architecture, that gives the storage vendors the ability to write MP software with insertion into the vmkernel
- Para-virtualization SCSI adapters
- SAS Para-Virt PCIe storage adapter
- Hardware spec written by VMware
- Provides functionality similar to bus logic, lsilogic and lsilogic SAS
- Supports MSI-X, PME, MSI capabilities in the device
- Drivers available for windows server 2003, 2008, and RHEL 5
- Key benefits---
- Lower overhead and higher CPU efficiency in I/O processing
- Higher throughput and low latency
- Better performance under high I/O conditions
- Better VM scalability (more VMs, vCPUs per host)
- Caveats-does not support boot disks with ESX 4
- VMdirectpath I/O (Experimental)
- Feature allows you to map a single HBA to a single VM. Prevents sharing of the HBA by more than a single VM.
- Allows VM's to directly access the underlying hardware devices
- vMotion, hardware independence and sharing of physical I/O devices not supported to the that VM using VMdirectpath I/O
- Experimental support for the following i/O devices:
-Qlogic QLA25xx 8G FC
-Emulex LPe12000 8G FC
-LSI 3442e-R and 3801e (1068 chip based) 3 Gb SAS adapters
