IBM N Series, Руководство по оборудованию

Страница: 312 / 368

292
IBM System Storage N series Hardware Guide
Firmware updates for disks in RAID 4 volumes are performed disruptively upon controller
boot unless the disk firmware is removed from the
/etc/disk_fw directory beforehand.
RAID 4 volumes can be temporarily (or permanently) upgraded to RAID-DP to automatically
enable background firmware updates (excluding gateway models). This operation doubles the
RAID group size. Therefore, it requires sufficient spares to add one double-parity disk drive
for each RAID group in a volume. To convert a traditional volume from RAID 4 to RAID-DP,
complete the following steps:
1. Convert the volume to RAID-DP by running the vol options <volume> raidtype raid_dp
command. Wait for double-parity reconstruction to complete.
2. Perform the automatic background disk firmware NDU as usual, followed by the Data
ONTAP NDU if necessary.
3. If wanted, convert the volume back to RAID 4 by using the vol options <volume>
raidtype raid4 command. This operation takes effect immediately. As a result, the
double-parity drive is ejected from the RAID groups, and the RAID group size is halved.
Manual disk firmware upgrade
To upgrade disk firmware manually, you must download the most recent firmware from the
IBM Support site to the controller’s
/etc/disk_fw directory. The disk_fw_update command is
used to start the disk firmware upgrade. This operation is disruptive to disk drive I/O. It
downloads the firmware to both nodes in an HA pair configuration unless software disk
ownership is enabled. On systems that are configured with software disk ownership, the
firmware upgrade must be performed separately on each node individually in sequence.
Therefore, you must wait for the first node to complete before starting the second.
Disk firmware can be downloaded only when the cluster is enabled and both nodes can
communicate with each other. Do not perform any takeover or giveback actions until the
firmware upgrade is complete. Firmware download cannot be performed while in takeover
mode.
Upgrades on RAID 4 traditional volumes and aggregates take disk drives offline until
complete, which results in disruption to data services. Disk firmware upgrades for
nonmirrored RAID 4 traditional volumes or aggregates that you did not perform before system
NDU must complete disruptively before the new Data ONTAP version can finish booting.
Storage system services are not available until the disk firmware upgrade completes. If not
updated previously, other disk drives, including spares, are updated after boot by using
momentary disk offline.
20.4 ACP firmware NDU
The EXN3000 disk shelves have a built-in component on the shelf module that is an
out-of-band control path to assist with resiliency on the shelf itself. This alternative control
path (ACP) requires separate firmware than the shelf modules. The ACP firmware update
process is an NDU.
20.4.1 Upgrading ACP firmware non-disruptively
Non-disruptive upgrades are performed by downloading the most recent firmware from the
IBM Support site to the controller’s
/etc/acpp_fw
directory. Updates start automatically for
any eligible ACP. Data ONTAP polls approximately once every 10 minutes to detect new
firmware in the
/etc/acpp_fw
directory. An automatic NDU firmware update can occur from
new firmware being downloaded onto either node in the
/etc/acpp_fw
directory.