HP StorageWorks XP Remote Web Console Software - Manual (Page 28)

28 Continuous Access XP Journal
• Cache overflow watch: If journal cache remains full for the time period specified by the cache overflow
watch parameter, the journal group is suspended due to a failure. This parameter can be 0 to 90
seconds.
If the amount of data in journal cache in the primary disk array reaches the specified journal cache
capacity, disk array I/Os are delayed. If journal cache remains full for the time period specified by the
cache overflow watch parameter, the primary disk array suspends the affected journal groups due to a
failure.
If the amount of data in journal cache in the secondary disk array reaches the specified journal cache
capacity, the secondary disk array stores received journal data in the restore journal volume, and issues
the next read-journal command to the primary disk array. This prevents an increase in the cache usage
rate.
Journal group operations
Continuous Access XP Journal journal groups maintains update sequence consistency across a journal
group of volumes. The primary and secondary data volumes of the pairs in a journal group must be
located within one physical primary disk array and one physical secondary disk array (1-to-1
requirement).
When more than one data volume is updated, the order data volumes are updated is managed within the
journal group to which the data volumes belong. Data update consistency is maintained among paired
journal groups. Continuous Access XP Journal uses journal groups to maintain data consistency among
data volumes.
Continuous Access XP Journal provides the following group-based operations to simplify and expedite
disaster or failure recovery procedures:
• Group operations at the primary disk array
• Split all pairs in a journal group. See ”
Splitting mirrors (Pairsplit-r: suspending copy operations) ” on
page 103 for a description of the Suspend Range-Group pairsplit option.
• Resume all suspended pairs in a journal group. See ” Restoring mirrors (Pairresync: resuming copy
operations) ” on page 104 for a description of the Resume Range-Group pairresync option.
• Delete all pairs in a journal group. See ” Deleting data volumes from mirrors (Pairsplit-S: ending
copy operations) ” on page 105 for a description of the Delete Range-Group pairdelete option.
• Group operations at the secondary disk array
• Split (pairsplit-r) all pairs in a journal group. See ”
Splitting mirrors (Pairsplit-r: suspending copy
operations) ” on page 103 for a description of the Suspend Range-Group pairsplit-r option.
• Delete (pairsplit-S) all pairs in a journal group regardless of their consistency status. See ” Deleting
data volumes from mirrors (Pairsplit-S: ending copy operations) ” on page 105 for a description of
the Delete Range-Group pairsplit-S option.
Continuous Access XP Journal data volume pair status
Continuous Access XP Journal displays the pair status for each data volume of a specified CU Image (CUI)
in the connected local disk array. Table 4 lists and describes the Continuous Access XP Journal data
volume pair statuses. The primary disk array maintains the primary data volume’s status and can change
the pair status of the primary and secondary data volumes. The secondary disk array maintains the
secondary data volume’s status and can change the pair status of the primary and secondary data
volumes, but not the primary data volume. The primary disk array detects when the secondary disk array
changes the secondary data volume’s status (if the path status is normal) and changes the primary data
volume’s status accordingly. You can display detailed pair status information using Command View XP or
XP Remote Web Console (Continuous Access XP Journal Pairdisplay pane) or a UNIX/PC server host (RAID
Manager Pairdisplay command).
A volume that is not assigned to a Continuous Access XP Journal data volume pair has the status SMPL.
When a Continuous Access XP Journal data volume pair is started, the primary disk array changes the
status of the primary and secondary data volumes to COPY. When the initial copy operation is complete,
the primary disk array changes the status of both data volumes to PAIR. When a pair is suspended due to
an error condition, the primary disk array changes the status of the primary and secondary data volumes
to PSUE (if the path status is normal). When you split a Continuous Access XP Journal pair (pairsplit-r), the