#Veritas volume manager support manual#

This parameter is not documented in the Veritas Volume Manager Administrator's Guide or in the vxdmpadm(1M) manual page. This can delay the failover process, and result in I/O sluggishness or suppression of I/O activity during the retry period. # vxdmpadm settune dmp_retry_timeout= valueДo not set the value of the retry period too high. To change the value of dmp_retry_timeout, type: To display the current value of dmp_retry_timeout, type: The retry period allows a fabric reconfiguration, which is usually a transient condition, to not be seen as an error by DMP. During this period, attempting to register PGR keys through the secondary path to an array may fail with an error condition, such as unit attention or device reset, or the return of vendor-specific sense data. This parameter has no direct effect on I/O processing by DMP.Дisabling a switch port can trigger a fabric reconfiguration, which can take time to stabilize. The default value is 120 seconds, which is twice the value of the SCSI I/O timeout. The maximum time period for which DMP retries the SCSI-3 Persistent Reserve operation with A/P arrays can now be tuned by using the dmp_retry_timeout parameter with the vxdmpadm command. Retry period for Persistent Group Reservation (PGR).The values of most DMP tunables can now be set by using the vxdmpadm settune command. Faster failure detection and recovery by DMPДMP has been enhanced to allow the vxdmpadm command to be used to adjust the I/O error recovery and I/O throttling mechanisms on individual paths.There are a number of enhancements to the Dynamic Multipathing (DMP) feature: The SmartSync feature of VxVM has been extended to support Oracle Resilvering of files in Veritas File System. In addition, the time taken to register keys has been reduced.

The keys for disabled paths are removed so that node joins are not obstructed. Http///docs/282024 SCSI 3 PR enhancement in clusters For the latest information on clustered volume support, see the following TechNote: The clustering functionality of VxVM supports up to 16 nodes. In this release, up to three sites are supported in a Remote Mirror configuration. A mechanism is also provided to deal with disk group inconsistencies between storage at different sites (serial split-brain scenario) and to perform fire-drill testing of site failure and recovery. The preferred read policy for mirrors has been extended so that read requests are serviced locally if possible. Existing utilities such as vxdg and vxdisk have been enhanced to achieve data set completeness, and to maintain site consistency. The Remote Mirror feature retains the current usage model for campus cluster support (also known as stretch cluster) where a disk group can span storage from multiple sites. Remote Mirror - robust campus cluster support The time that nodes take to join a cluster has been reduced by allowing concurrent joins. The time taken to start the Veritas Volume Manager configuration daemon, vxconfigd, to discover new devices, to initialize dynamic multipathing, and to import disk groups has been significantly reduced. The changed behavior applies to both private and cluster-shared disk groups. As using the -f option to forcibly import an incomplete disk group counts as a successful import, the disk group can subsequently be imported without specifying this option. With this release, a disk group can be imported successfully if all the disks are accessible that were visible when the disk group was last imported successfully. In such cases, the -f option had to be specified to forcibly import the disk group. In previous releases, the vxdg import command could not import a disk group if some disks in the disk group were unusable or inaccessible. You do not need to change the private region size of existing disk groups to the new default value. The default size of the private region has been increased from 1 MB to 32 MB. This release of Veritas Volume Manager includes the following new features and enhancements.

Veritas volume manager support