HP XP7 Storage - Manual (Page 163)

Figure 53 Sidefile quantity limit
Sidefile area: Sidefile area = 30% to 70% of cache as set on Remote Web Console (default sidefile
= 50% for XP24000/XP20000 Disk Array, XP1024/XP128 Disk Array; 40% for XP12000 Disk
Array/XP10000 Disk Array).
Write I/O control via the high-water mark (HWM): When the quantity of data in the sidefile reaches
30% of cache, the Continuous Access Asynchronous pair status is HWM of PAIR state, and the
host write I/Os receive a delayed response in the range of 0.5 seconds to 4 seconds. Following
is an arithmetic expression of the HWM at 100% of a sidefile space:
HWM(%) = High water mark(%) / Sidefile threshold (30 to 70) * 100
Write I/O control via the sidefile threshold: When the quantity of data in the sidefile occupies the
maximum defined sidefile area, host write I/Os are delayed until there is enough sidefile space
to store the next new write data. The copy pending timeout group option is defined using Remote
Web Console and specifies the maximum delay between the M-VOL update and the corresponding
R-VOL update. The range for the copy pending timeout option is 1-255 seconds (600 seconds for
Continuous Access Journal), and default value is 90 seconds (60 seconds for Cnt Ac-J). If the timeout
occurs during this wait state, the pair status changes from PAIR to PSUS (sidefile full), and host
write I/Os continue with updates being managed by the cylinder bitmap. Important: The copy
pending timeout value should be less than the I/O timeout value of the host system.
Continuous Access Asynchronous transition states and sidefile control
Continuous Access Asynchronous volumes have special states for sidefile control during status
transitions.
Table 36 (page 164) shows the transition states for Continuous Access Synchronous and
Continuous Access Asynchronous volumes.
The suspending and deleting states are temporary internal states within the RAID storage system.
RAID Manager cannot detect these transition states, because these states are reported on the
previous state of the storage system. These states are therefore concealed inside the pairsplit
command. After the pairsplit command is accepted, host write I/Os for the P-VOL are managed
by the cylinder bitmap (normal), non-transmitted data remaining in the P-VOL's FIFO queue is
transferred to the S-VOL's FIFO queue, and the pair status is then set to PSUS [SMPL] state when
all data in the P-VOL's FIFO queue has been transmitted.
PFUL. If the quantity of data in sidefile cache exceeds 30% of cache storage, the internal status of
the RAID storage system is PFUL, and host write I/Os receive delayed response in the range of
0.5 seconds (minimum) to 4 seconds (maximum).
PFUS. If the quantity of data in sidefile cache exceeds the user-defined sidefile area (30%-70%),
then host write I/Os must wait for enough sidefile space to become available for storing the next
new write data. If a copy pending timeout occurs during this waiting state, then the pair status
changes from PAIR to PFUS, host write I/Os are accepted, and write data is managed by bitmap.
The RAID Manager software can detect and report the PFUL and PFUS states as follows:
•
As a return code of the pairvolchk command
•
As the status code displayed to code item by the pairmon command
•
As the paired status displayed to status item using -fc option of pairdisplay command
Continuous Access Synchronous, Business Copy, and Continuous Access Journal operations 163