Storage Expiry does not reclaim space as expected on Dell Elastic Cloud Storage (ECS) partition that emulate Centera partitions. 

book

Article ID: 100047095

calendar_today

Updated On:

Description

Error Message

There is no error associated with this issue. However, it can be witnessed where regular Storage Expiry runs show large amounts of data being expired but little or no space appears to be reclaimed on the (ECS) device in comparison. 

Dtrace of StorageDelete will show the following:

Call to storage to delete a specific clip id:  

[6184]    (StorageDelete)    <7868>    EV:M    CVaultStoreEMCCentera::ClipDeleteCenteraCollectionDelete -- Calling FPClip_Delete (Local) -- Clip-Id: 6J0P7JJDE8ES6e8EUGUL1A0D83CG417CDDLCUG0GD2IBCHUR69LKS

 

The StorageDelete will get the return that the clip has been deleted:

[6184]    (StorageDelete)    <7868>    EV:M    CVaultStoreEMCCentera::ClipDeleteCenteraCollectionDelete -- Returned from FPClip_Delete (Local) -- Clip-Id: 6J0P7JJDE8ES6e8EUGUL1A0D83CG417CDDLCUG0GD2IBCHUR69LKS
 

At this point Enterprise Vault will proceed with removing associated SQL and Index entries. 

Important to Note:

When the device returns that it has deleted the clip it is not actually deleting the entire clip. Instead it will delete the pointer to the clip known as the Clip Descriptor File (CDF) and leave the associate Blob(s) on the device. These Blobs make up a majority of the entire Clip. A process on the device known as Garbage Collection deletes the rest of Clip at a later time. It is the Garbage Collection process on the (ECS) that is not reclaiming space in a timely manner as it should or would be expected. 

Cause

Per (ECS) 3.4 release notes Issue # STORAGE-24171 "Chunks with 48B padding would be forced to put through Garbage Collection verification impacting overall Repo garbage reclamation rate." has been resolved with the release of (ECS) Revision 3.4.x. 

Please see ECS 3.4 Release Notes below.

Note: A separate login is required. 

https://support.emc.com/docu95769_ECS-3.4-Release-Notes.pdf?language=en_US

Resolution

After upgrading to (ECS) Revision 3.4.x customer experiencing this issue have seen drastic improvements in space reclamation. 

For more information regarding this issue please contact Dell (ECS) support. 

Issue/Introduction

Storage Expiry does not reclaim space as expected on Dell Elastic Cloud Storage (ECS) partition that emulate Centera partitions. This issue has been observed on data that has been migrated from EMC Centera devices to (ECS) partitions that emulate Centera devices. This issue has been observed (ECS) Revisions 3.2.x.
Powered by Wolken Software