EMC recently released Virtual Storage Integrator 6.5 and I wanted to share some of the new features that were added. If you are using EMC storage (especially XtremIO) then you need to be using VSI!
This release includes the following new features:
- Support for VMware vSphere 6.0
- Support for setting multipathing policies from the VSI GUI using native multipathing and PowerPath®/VE
- Note: The PowerPath multipathing setting is supported in vSphere 6.0, only.
- Support for selectively disabling snapshots during datastore protection in AppSync
- Support for the following functions on EMC VNXe3200 storage:
- Extend datastores on a thick or thin LUNs and NFS file systems
- Enable and disable deduplication and compression on VNXe3200 datastores and virtual machines
- Note: The default file system size for VNXe3200 version 3.1 is 32 bit. For VNXe3200 version 3.1.1, the default file system size is 64 bit. For version 3.1 you can enable and disable file compression and deduplication. These functions are not available for a 64-bit file system.
- Support for client registration on EMC VMAX3 storage
- Support for the following features on VPLEX storage:
- View the properties of the VPLEX storage system, VMFS datastores, and RDM volumes
- Provision VMFS datastores and RDM volumes to virtual machines
XtremIO 3.0.3-11 was just released and below is a list of the fixes. These can be found in the release notes for 3.0.3-11.
General Fixed Issues
- SSD status check has a wrong timeout and may, in some rare cases of SSD failure, cause an unexpected service stop.
- In some rare cases, resetting a new SSD in the DAE causes an unexpected service stop.
- SSD insertion is detected by only one Storage Controller, and in some rare cases causes an unexpected service stop.
- Deleting a high number of snapshots or creating and deleting a snapshot of the same volume at the same time may cause an internal module restart that does not cause any high availability event.
- The System Manager module fails and restarts when the add-initiator command is sent to a non-existing Initiator Group.
- OpenSSL has been upgraded to openssl-1.0.1e-30.
- OpenSSH has been upgraded to openssh-5.3p1-104.
- HTTP TRACE method is disabled.
- Replacing an SSD without following the correct procedure causes the SSD replacement to fail with “SSD not initialized” and “SSD is not configured in any XDP group” alerts. The replaced SSD is not cleared from the repository.
- Fabric WWN is displayed as an initiator.
- In some rare cases a high volume of deletion I/O causes an increase in the latency.
- SCSI registration preempt with type change can cause an unexpected module restart.
- A “sense data” response is sent during registration conflict, even when the SCSI status is not “Check Condition”, which causes Emulex HBA to ignore the registration conflict.
Upgrade and Cluster Hardware Upgrade Fixed Issues
- The create-snapshot command may receive a busy response if the cluster is going through OS upgrade.
- In some rare cases, issuing the read-reservation SCSI command after NDU to a registered path that was registered before the NDU, can cause the cluster’s state to appear as “unknown”. The cluster automatically recovers with no user intervention.
- SCSI reservations that are made during NDU cannot be released.
High Availability Fixed Issues
- In some rare cases, losing power on one InfiniBand Switch causes an unexpected service stop.
UI and XMS Fixed Issues
- XtremIO SNMP trap OID contains two redundant digits.
- Entering unsupported characters while modifying a password may cause the password to be unusable.
- SSH key based users are removed during the cluster maintenance.
- The severity of sys_ud_ssd_space_very_limited alert (Cluster capacity at 90%) has been changed to “Major” severity.
- The thresholds for free physical space alerts have been changed to:
- Limited Free Space threshold : 15% (85% used)
- Very Limited Free space: 10% (90% used)
- No free space: 3% (97% used)
- Using RESTful API with External User (LDAP) can cause XMS unavailability. Using a local account for RESTful API prevents the issue.
- XMS recovery fails when the configuration has SSH key based users with public key longer than 2048 bytes.
- OpenLDAP authentication fails when “uniquememeber” is used as part of the group membership search.
- RESTful API requests to /api/json do not receive a proper response and time out.
- LDAP authentication fails when the user name contains special characters.
- Added the NAA identifier to the show-volumes command.
- The modify-cluster-passwords have been changed to allow changing the “xinstall” password. For changing the cluster’s passwords, refer to EMC KB 183472.
- Added command to clear SCSI reservations.
- Added a warning message and confirmation request to the clear-volumereservation command.
- A warning message and additional confirmation were added to the freshinstall command to prevent the installation of an existing cluster.
- Warning messages were added to the stop-cluster command.