Background Tasks; Error Handling - Intel SE7520AF2 Technical Product Specification

Integrated Intel® RAID Controller SROMBU42E
Write Back - I/O completion is signaled when data is transferred to cache.
Cache Policy
Direct I/O - When possible, no cache is involved for both reads and writes. The data
transfers will be directly from host to disk and from disk to host.
Cached I/O - All reads will first look at cache. If a cache hit occurs, the data will be read
from cache; if not, the data will be read from disk and the read data will be buffered into
cache. All writes to drive are also written to cache.
Read Policy
No Read Ahead - Provides no read ahead for the logical drive having this option turned
on.
Read Ahead -Additional consecutive stripes/lines are read and buffered into cache.
Adaptive - The read-ahead will be automatically turned on and off depending upon
whether the disk is accessed for sequential reads or random reads.
4.3.5

Background Tasks

Rebuilding a failed drive is performed in the background. The rebuild rate is tunable from 0-
100%.
The rebuild rate controls the amount of system resources allocated to the rebuild
process.
It is not recommended to increase the rebuild rate over 50% as a higher rebuild rate can
result in the operating system requests not being serviced in a timely fashion and
causing an operating system panic.
A consistency check validates parity or mirror integrity and can fix a parity or mirror error, if
necessary. Consistency checks use the same rate as a rebuild.
Background initialization is a background check of consistency. It has the same functionality as
the check consistency option but is automatic and can be canceled only temporarily. If it is
canceled, it will start again in five minutes or less. Background initialization is only performed on
redundant volumes.
RAID level migration and online capacity expansion are also completed in the background.
4.3.6

Error Handling

Most commands are retried four or more times. The firmware is programmed to provide the best
effort to recognize an error and recover from it if possible. Any failures are logged and stored in
NVRAM. Operating system-based errors are viewable from the event viewer in Web Console.
RAID-related errors can be reported by the hard drive firmware, SAF-TE controller, or the RAID
controller firmware. These errors may be reported to the operating system through RAID
management software, through SMART monitoring, or through CIM management. Some errors
may also be reported by the SAF-TE controller and logged in the system event log (SEL). In
addition, access errors may be reported by the operating system. Depending on the RAID
88
Intel order number C77866-003
Intel® Server Board SE7520AF2 TPS
Revision 1.2