Optimising housekeeping services on HPE Storeonce

[AWR]

Hi,

Please can you advise me? We have an HPE Storeonce 5650. Is there anyway to optimise the housekeeping services on the HPE Storeonce, for example if we delete a big backup, e.g. 31TB, the space is only reclaimed much later when the house keeping services run. It also appears that sometimes the housekeeping services do not run due to capacity constraints. 

Appreciate any inputs and advice.

Kind regards

Andrew Rycroft

[support_s]

System recommended content:

1. HPE StoreOnce Software Version 4.3.2 Release Notes | Fixes

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

[ArendL]

Dear Andrew,

You post does not show the software version you are using.

There's a couple imrovements in 4.3.5, 4.3.6 and 4.3.7.

31 TB is quit a lot, but, is that user bytes or disk bytes?.

When is the retention period, is that perhaps 7 or 14 days only?.

Regards,
Arend

[Mohammed_I · Accepted solution]

Hi Andrew Rycroft

- Housekeeping is an important process to maximize the deduplication efficiency of the appliance. Ensure that the housekeeping process has enough time to complete by not overloading the device. For example, running backup, restore, replication, and StoreOnce Catalyst Store operations with no break (that is, 24 hours a day) can result in housekeeping never completing. Use Housekeeping statistics to determine if the system is appropriately loaded. If the HPE Confidential Best practices 260 housekeeping pending queue is steadily increasing, you should decrease the load on the device to allow for more housekeeping processing.

- Batch deleting a large amount of data at one time causes the pending housekeeping queue to spike higher than normal. If you are experiencing an already high pending housekeeping queue, then reduce the amount of data deletion per batch, and delete data more frequently.

- Having many VTL libraries/NAS shares/Catalyst stores that each have a common data type can lead to better Housekeeping performance versus having one large store that contains a mixture of data type

Regards, 

Mohammed