Navigation:  Features > Filesystem Disk Management Support

Introduction

The ImageServer has a number of features that support disk management of its filesystems.  The Rules Engine sets when studies are eligible for the various disk management options.  The Service Scheduling page's Filesystem Watermark Check item is used to do the actual watermark check for a filesystem and scheduling of operations in the WorkQueue for the filesystem to reach the low watermark.

When a rule is defined for a study, it defines when a study is eligible for disk management.  The rule also defines the relative order compared to other studies for the various disk management operations.  The study will not be tier migrated, deleted, or purged unless the filesystem it resides on is above a high watermark, and the Service Scheduling page's Filesystem Watermark Check determines the study is the next eligible study available for the operation.

The ImageServer can be configured to run as a temporary cache or as an archive where studies are stored "nearline" on an HSM based archive. The mode of operation is determined by the disk management related rules defined. With its default installation, the ImageServer is configured to run as a temporary cache, with no studies being archived.

Note that compression of online studies is treated independently from disk management. Studies are automatically compressed as specified by the rules engine independent of the watermark of the filesystem the study is stored on. See the compression section for details on compression. Also, see the Rules Engine section for a more thorough explanation of the rules engine.

The following sections describe in further detail how the ImageServer does disk management and includes some example rules.

Configuring Disk Management Rules

Disk Management rules are configured through the Rules Engine Configuration screen.

Disk Management Rule Types

There are three types of disk management rules in use for the ImageServer:

In addition to telling when a study is eligible for disk management, the time also specifies the relative order that disk management will occur on studies as compared to other studies on the system. The ImageServer sorts all eligible studies for disk management by the scheduled time, processing the first eligible studies first.

There are three different times when rules are run against the study. The Study Processed apply time is when the study is initially processed by the server. These rules are also executed when the Service Scheduling Reprocess Studies queue item is run. The Study Archived apply time is run after a study has been archived. The Study Restored apply time is run after a nearline/offline study has been restored back to online status from an archive.

Temporary Cache vs. Archiving

As mentioned above, the ImageServer is configured as a temporary cache by default. This is caused by a default rule Study Delete rule being defined. This default rule must be disabled in order for the ImageServer to function as an archive.

Note, however, that the ImageServer can function in a mixed mode, where some studies will be deleted and others archived. Study Delete rules can be defined such that specific studies that match a condition are deleted, while all other studies are archived.

Online Retention rules specify how long an archived study should be kept online. Online retention rules must be defined for both when a study is archived, and when it is restored from offline. Note that typically an online retention when a study is processed might have a time in relation to the study date of the study, but an online retention rule run at restore time will use an elapsed time from when the study is restored to specify a minimum amount of time for the study to stay online.

Tier Migration

You may have noticed that a rule type is defined for migrating off of tier 1, but not for migration off of other tiers. When a study migrates off the first tier of storage, the ImageServer simply uses a timestamp at the time of migration to tell when the study is eligible for migration to subsequent tiers. This means that studies will migrate off of lower level tiers in the same order they migrated off of tier 1 storage.

The ImageServer defines a default rule for tier migration. If you only have filesystems defined for tier 1, these rules will still be applied and the ImageServer will still keep track of when a study is eligible to be migrated from tier 1. This does make it possible to introduce filesystems assigned to lower tiers at any time.

Scheduling of Disk Management

The Service Scheduling mechanism is used for checking if a filesystem is above the configured high watermark for the filesystem. A service scheduling queue entry is created for each filesystem. The queue entry schedules itself to periodically check if its filesystem is above the high watermark. The service scheduling queue entry can be disabled if the user desires to disable disk management temporarily or permanently on a filesystem.

Order of Operations for Disk Management

When a filesystem exceeds the high watermark, the ImageServer will first check to see if any studies are eligible to be migrated to a lower level tier filesystem. If the ImageServer is only configured for a single tier of storage, this step will be skipped. Next, if necessary, the ImageServer will check if any studies are eligible for deletion. Finally, the ImageServer will check if any study is eligible for purging from the filesystem. Purging will only occur when a study has been archived.