Backup Exec 20 Deduplication Improvements

[tape_monkey1]

Are there any details regarding the improvements to the Deduplication Engine from Backup Exec 15 to Backup Exec 20?  the statement "Backup Exec users can get improved speed, scale, and stability while backing up to the Backup Exec deduplication folder." from the BE 20.1 release notes is pretty vague.  Is there any published data on the improvements?

[Gurvinder]

64 TB max size is maintained. Same concepts about client, server and appliance side dedupe exists. DD version was 7 in BE 15, BE 20 uses DD version 10. The PostgreSQL database to house the segments and data relationship is gone and a flat file database is now used. There is a mutithread service to improve the performance. It is more stable in terms of handling of internal corruption or keeping a check in integrity of data in dedupe. Performance in terms of job rate is better than earlier.

[deejerydoo]

Since upgrading to v20 I am seeing no deduplication improvements whatsoever, over the previous v15 system we had.

All I am seeing now is a server that has gone to 24/7 high utilisation, just doing a never ending task of checking the deduplication store, only pausing this when backup jobs run.

Improved compression ratios - No.

Higher throughput for backup, verification, cataloguing, duplication to tape jobs - No.

Improved deduplication store maintenance - No.

This was an in place upgrade, so the hardware has not changed. The hardware is a HPE Proliant DL180 G6 2 x E5540 quad core @ 2.53GHz, 24GB RAM, Win2K8 with a 12 disk SATA DAS through a Smart Array P411 array controller. It's managing a 6.4TB deduplication store (about 50TB of undeduplicated data). I know this hardware is old but it is getting no improvements where you state it will, with the new software.

[Gurvinder]

Can you attach the Adam.log from BE install path\logs

What are you predominantly Backup up -

Can you comment on the observations post the upgrade for dedupe -

1) how is the verify job rate ? Have you seen a dip or they have been the same compared to earlier BE version.

2) how is the dedupe ratio ? You can check
It in dedupe storage details in storage tab of BE. Normally the dedupe ratio remains the same.

3) the improvements are w.r.t the continuous check which runs on dedupe drive and this version deals better with any corruption. There are mechanism to self heal. This version also removes PostgreSQL and uses Flat file database which helps for improved transaction fetches etc. you can refer NBU 7.6 ( at a high level the dedupe mechanisms are common)

https://vox.veritas.com/qsuop67736/attachments/qsuop67736/Articles-Backup-and-Recovery/738/1/NetBack...