IBM totalstorage 326, Справочник пользователя

Страница: 81 / 240

concurrently on the drive, given the amount of file-write activity to the drive. PSM
automatically takes action to prevent the cache file from overflowing, because if
that occurred, PSM would be forced to automatically delete all persistent images
on the drive (when it cannot keep track of changes made to the drive, it cannot
maintain a valid persistent image).
PSM takes the following actions as the cache file usage approaches a full
condition:
– When the cache file usage exceeds the warning threshold (configured in the
PSM Volumes panel for the drive; the default value is 80 percent), PSM
generates a warning message to the system event log (viewable through the
Windows 2000 Event Viewer in the IBM NAS Admin console), and to the alert
log in the Microsoft Windows 2000 for Network Attached Storage user
interface. The name of the source for the message is
psman5 . Additionally,
while the cache file usage is above the warning threshold, PSM prohibits any
attempt to create a new persistent image, and logs error messages (to the
system log and alert log). The text of the error message that is logged in the
system event log (from
psman5 ) is “A persistent image could not be created
due to error 0xe000102b”.
– When the cache file usage exceeds the automatic deletion threshold (also
configured in the PSM Volumes panel for the drive; the default value is 90
percent), PSM automatically selects a persistent image on the volume and
deletes it to reduce the cache file usage. It selects the persistent image with
the lowest retention value (as described in “Persistent Images” on page 62). If
more than one persistent image has the same (lowest) retention value, then
the oldest image will be selected for deletion. If this deletion does not reduce
the cache file usage below the automatic deletion threshold, then it will
continue to select and delete persistent images until the cache file usage is
reduced below the automatic deletion threshold. For each deletion, PSM
generates an error message to the system event log and to the Windows
2000 for Network Attached Storage alert log indicating that a persistent image
was deleted.
You should periodically check the system event log or Windows 2000 for Network
Attached Storage alert log to ensure that the cache file usage is not consistently
high, forcing existing persistent images to be deleted and preventing new
persistent images from being created. If the cache file usage is high, you can
increase the size of the cache file using the PSM Volumes page. However,
because dynamic cache file resizing is not supported in this release, you must
delete all persistent images currently on that volume first.
v
When a shared volume is failed over from one engine in the Model 326 to the
other engine, the persistent images for that volume move with the volume. The
Persistent Images panel on a particular engine will display only those persistent
images which are on volumes that the engine owns at a point in time. If
persistent images are scheduled for a volume, on a particular engine, a
scheduled persistent image is created only as long as that engine owns the
volume at the time the scheduled persistent image is to occur.
To ensure that a scheduled persistent image will take place regardless of which
engine owns the volume, you must do the following:
1. Use the Schedules panel to create the schedule on the engine that currently
owns the volume.
2. Use the Cluster Administrator to move the disk group that contains the
volume to the other engine. You can only create or edit a schedule for a
volume on the engine that currently owns the volume. If an engine does not
Chapter 6. Managing and protecting the network and storage
69