Monday, July 11, 2022

Backup and Recovery materials

Data Domain DD6400 - Why? Architecture Costs - video: 
Data Domain DD6400 - Why? Architecture Costs - article and presentation: 
Data Domain DD6400 - Why? Architecture Costs - article and presentation:

Why new Data Domain models? DD6900 / DD9400 / DD990


Cyber Security / Cyber Bunker - detailed explanation
Presentation #1:
Presentation #2:

CyberSense - How to be sure my Data Center is clean
Video - English: 
Video - Polish:

Fast backup of VMware - Transparent Snapshots - video:
Fast backup of VMware - Transparent Snapshots - article:
Fast backup of VMware - Transparent Snapshots -  presentation: 

Fast backup of files - PPDM and file shares - video:
Fast backup of files - PPDM and file shares - article and presentation:
Fast backup of files - PPDM and file shares - presentation:

Article describing vProtect:

Performance in backup

NetWorker presentations:
NetWorker - why, architecture, value: 
NetWorker for VMware:

What is new in Data Domain 7.2?

Full presentation packages

PowerProtect Data Manager presentation packages:
PPDM - Why? Architecture, Demo
PPDM - Installation / Configuration - Step by step guide:

Sunday, July 10, 2022

Data Domain DD6400

Video: Data Domain DD6400 - Why? Architecture, Costs - Security and performance in backup / recovery

Data Domain is well know as unbeatable security / performance / easy of use.
Can new Data Model model - DD64400 - can add anything?
Why the values of Data Domain matters?
Have a look on short /detailed video!

Direct link to the video:

Detailed description of topics covered in the video: 
0:05 Customer story
- customer encrypted wants to protect for future using Data Domain
 - New DD6400 provides security, performance, scalability

1:02 Agenda

1:15 Backup software and backup media

1:25 - a lot of backup media - tape / disk. why to use Data Domain?

1:39 All Data Domain models

2:03 Performance
 - we write 1%
 - Data Domain does not write data to backup media during backup

2:53 - Performance numbers: 14TB/h for MSSQL environment

3:03 - Oracle speed 5TB/h
 - 1.5GB/s
 - limitation was production storage

3:40 Parallel backup of all items in production environment
 - Virtual machines / databases
 - 270 streams
 - comparison to concert and toilet
 - backup environment - one of the best to perform parallelly

4:37 Additional speed because of Storage Pools

5:58 Why additional speed for backup?
 - Data Domain does not write data during backup
 - We need the speed for recovery
 - There are other mechanisms as defragmentation

7:10 Data Domain speed from de-duplication
 - variable length de-duplication
 - example with 3 Virtual Machines built from 3 blocks - few space required
 - new yellow block and compression required
 - DD6300 - compression by main processor
 - DD6400 - compression card that performs the compression
 - processor is free for replication, cleaning, etc.
 - blocks are smaller - better compression
 - faster backup / restore but also the whole data domain is faster!
 - all new Data Domains: DD6400 / DD6900 / DD9400 / DD990 have the compression card

10:00 Starting backups directly from Data Domain!
 - Instant Access
 - We want to restore immediately Virtual Machine
 - We want Data Domain to be primary storage
 - Can we start Virtual Machine from tape? IT would be like asking kid to empty the trash
 - SSD disk speed

11:50 Revolution in backup
 - normally if something fails - we are offline until backup will restore data - long time
 - now if production fails, backup becomes the production
 - example with customer who was running 75 VMs directly from Data Domain - failure of firmware upgrade in disk array

12:30 Security
 - checking backups on the fly
 - whatever backup software we have - all data are red and checked if correct
 - guarantee of recovery correct data

 - Storage Pool keeps a lot of additional data 
 - this architecture allows to fail more disks than in RAID-6

15:15 Separated disks for Operating System and data
 - we can reinstall the Operating System and connect disks with backups

16:25 Logically separated Operating System and data - different file systems

16:48 Replication between Data Domains
 - whole site can be damaged
 - we transfer 1% of the data - we can restore 100% of data
 - case study - 29 locations replicated to central site
 - customer wants only Data Domain for Disaster Recovery base on the great experience
 - almost no transfer, no network usage, reliability

17:48 Hacker / ransomware protection
 - compliance - locking backups for any period
 - no one can destroy my backup

18:12 Demo of Data Domain compliance
 - we setup lock for 3 months
 - for whatever backup software we have secure backups against ransomware / hacker attack 

20:49 Cyber Bunker
 - Data Domain is base for Cyber Bunker - gold copy that neither hacker/ransomware can destroy

21:18 Boost - secure protocol
 - Boost protects us against ransomware
 - Boost is protected by login/password that ransomware does not know

21:40 Security architecture of DD6900 / DD9400 / DD9900
 - separated disks for Operating System / Data
 - separated logic

22:00 Retention Lock / Compliance with Veeam Backup Software
 - Veeam does not force automatically Compliance
 - Cyber Bunker is solution for Veeam
 - Lock is than based on Data Domain in Cyber Bunker
 - Possible to make lock on production - snapshot which is locked

23:17 Setup lock can be on destination Data Domain only

23:30 We can make incremental / Synthetic FULL backups with compliance mode

23:42 Number of Virtual Machines that we can start DD6400
 - 20 Virtual Machines can be started from DD6400
 - with more than performance will be smaller

24:05 If we replace disks with Operating system what about configuration?
 - data on Data Domain are self-described
 - we can install new Operating system and attach to Data Domain

24:52 NetWorker - how to setup retention lock on Data Domain set-up from NetWorker level
 - for any backup software we can setup retention lock using Cyber Bunker

26:00 Capacity / Scaling
 - DD6400 starts from 8TB net and scales up to 172TB
 - we can increase capacity in 4TB increments
 - other models starts from much bigger capacity apart from DD3300
 - DD6400 - we start from 8TB and we increase capacity with just entering license - 4TB increments
 - increase capacity with entering licensing is possible up to 32TB 
 - 40TB of net capacity requires adding hardware - new shelf
 - than we increase capacity again but adding just license
 - we can say DD6400 provides "capacity on demand"
 - ordering DD6400 we just provide net capacity with 4TB increments

29:05 Data Domain - best market de-duplication
 - small variable block with global de-duplication
 - it writes almost no data
 - if we backup guest level and image level the second backup writes almost no data -> global de-duplication
 - source de-duplication / Virtual Synthetic
 - 2-3 less space than any other competition because of global de-dup / variable & small block de-duplication

29:57 Customer story and Data Domain global de-duplication
 - 9TB VMware 
 - first backup occupied 0.6TB (600GB) - almost nothing
 - 1:15 de-dup with just 1st backup
 - many products have issue to overcome 1:8 de-duplication with over 30 days retention
 - here it was just after 1st backup
 - Data Domain has 1:100 to 1:300 de-duplication with 30 days retention
 - this makes Data Domain cheap!

30:52 Small customer - 300GB of SAP/Oracle
 - De-duplication after 4 backups was 1:15  - great one!

31:12 How much Data Domain do I need for my backups?
 - For 10TB / 30 days retention usually we need 8TB
 - rule of thumb
 - huge small amount

31:37 Tool for sizing Data Domain - Sizer

31:52 Compression card
 - we can even further decrease the cost!
 - example with VMs built from 3 blocks
 - if Data Domain sees a new block he must write to disk
 - Data Domain will compress new block further
 - DD6300 - lz algorithm for new blcosk - lz not very effective but very little impact on DD processor
 - DD6400 - compression of new blocks is made by compression card
 - much faster
 - better compression, much less space is required(gz or gzfast)

33:40 How less space DD6400 takes over DD6300?
 - Migrating from DD63300 100TB to DD6400 - DD6400 will require only 80TB space
 - reason -> lz compression of new blocks vs gz/gzfas on DD6400
 - these number are from worldwide statistics
 - detailed stats for files / databases
 - compression card is in DD6400 / DD6900 / DD9400 / DD9900

35:45 Customer example tat migrated from DD4200 (no compression card) to DD6900 (with compression card
 - Space dropped from 27TB (DD4200) to 20TB (DD6900)

36:05 Summery - DD6400 with the same net capacity is cheaper than DD6300
 - similar price
 - less storage required

36:42 - Parameters of DD6400
 - part of new models family - DD6900 / DD9400 / DD9900
 - DD6400 capacity from 8TB to 172TB net with increments of 4TB
 - capacity of other models
 - DD6400 uses 8TB drives
 - DD6400 has 270 simultaneous backups -> huge important
 - streams of other models
 - DD6400 is 2U / 3U shelf - max 2 shelves
 - DD6400 built in Ethernet card 10Gb -> SFP/BasetT to choose
 - DD6400 allows for 3 additional Ethernet cards - the same as above  or 25Gb
 - DD6400 allows for 1 additional FC card

39:35 Additional materials
Backup and Recovery - all materials:
BackupRecoveryGuy: Backup and Recovery materials
DD6400 video: 
DD6400 article: 
New generation of Data Domains DD6900 / DD9400 / DD9900
What is new in Data Domain 7.2?
Performance in backup:
Fast backup of files - video:
Fast backup of files - article and presentation:

40:00 Summary
 - please consider DD6400 -> speed, security,  flexibility, cost-effective, flexibility
 - the above give Data Domain 80% of market
 - please check!

Presentation from the event: 

Only successful restores!

Saturday, July 9, 2022

Fast backup of files

Video: Fast backup of file shares - No load, ransomware protection, fast restore

Millions of files? Huge number of TBs?
Slow backup? Issue with restore?

There is a solution! You can backup any file share fast, secure and with easy restore!
See summary of what PowerProtect Data Manager offers for backup of CIFS/NFS - any NAS shares!

  • Huge fast - backup every hour! 
  • Hardware agnostic
  • Fast restore

Direct link to the video:

Detailed description of topics covered in the video:
0:05 - Backup of files can be pain:
 - load
 - time

0:30 - Agenda

0:50 - Why do we talk about backup of files?
 - we talk about CIFS/NFS that is tough to backup
 - no agent possible as for Windows NTFS use case
 - we have millions of files / hundreds of TB
 - backup is slow - even days
 - restore is slow, restore point is long time ago

2:58 - We talk about PowerProtect Data Manager
 - can backup fast CIFS / NFS shares from any vendor

3:29 - Architecture of the backup file share
 - we have FULL backup by only reading what has changed since last backup
 - backup is huge fast because we read only delta
 - PPDM performs only FULL backups - though reads only delta
 - no incremental / differential backups - only FULLs 
 - huge important because FULL backups provides fastest restore and possibility for granular restore

6:40 - Value
 - no load of protected system / protected system does not feel the backup
 - no load of network - protected share can be far from media
 - backup is huge fast / I can perform backups frequently 

8:12 - What does t mean that every backup is full but we read delta?
 - if 2 files have changed, we will read just them
 -  we have a full backup
 - this is magic and incredibly fantastic

10:12 - PPDM uses BOOST protocol

10:25 - It is magic and fantastic!
 - during couple of minutes we can have a full backup

 10:45 - PowerProtect Data Manager demo
 - PPDM just controls backups - does not take part in data movement
 - we add a new policy for CIFS backups
 - we choose share
 - we can choose synthetic full - full with reading delta

12:35 - Automated dividing for streams / slices
 - dynamic slicing
 - every 200 GB is single stream
 - 3 streams for 600 GB data
 - every 1 million of files is single stream
 - 5 streams for 5 millions of slices

14:15 - demo - setup maximum number of streams for particular share
 - we do not want to use to many streams for particular streams
 - limiting streams for a large number of TBs

16:15 - The backup is controlled by proxy
- PPDM controls backups from aside 
- proxy is installed automatically from PPDM
- I can have many proxy - every of them is 24 streams

17:26 - This is functionality for all possible CIFS/NFS shares
 - any CIFS any NFS on whatever storage
 - hardware agnostics
 - we can enjoy huge speed
 - it is just software / proxy
 - proxy is just Virtual Machine
 - we can add a new proxy / VM to add streams
 - 24 streams per proxy

19:12 - How many Data Domain systems are supported?
 - as many as we want
 - we can use bigger one - DD9900 - almost 2000 streams

19:38 - No agent required
 - agentless 100%
 - we do not install anything - also on the backed up share
 - adding a new share - we will se no agent is required
 - showing how simple is adding a new share to PowerProtect Data Manager

21:20 - Proxy binaries are within PPDM
 - demo how easy is to install the proxy

21:50 - Only first backup is reading whole data
 - every following backup is reading change but we have a full backup

22:07 - Proxies are deployed manually
 - we shall have more calculate how many streams we require
 - if we need 50 streams probably we need 3 proxies (x24 - 72 streams) - to have some buffer

22:47 - What do we get? Summary:
 - Speed 
 - Frequent backups
 - All backups are FULL
 - Immediate restore
 - Granular restore

23:02 - Protection against ransomware / hacker attack
 - we can setup retention lock
 - demo - setup retention lock live in PPDM
 - production can be deleted but our backup will survive

24:17 - Self-backup infrastructure
 - demo of self-backup infrastructure
 - when we create policy we need to manually say which shares we want to backup
 - but we want that PPDM automatically add new shares connected with PPDM
 - showing dynamic filters live
 - all shares that in the name contain k8s must be automatically backed up
 - we see that now PPDM will backup 4 shares having k8s within name
 - we can also say that big shares go to one policy / small shares go to another policy
 - all backups from some Isilon shall go somewhere
 - Self backup infrastructure is available only for Unity / PowerStore / PowerScale

27:35 - What is Self-backup infrastructure?
 - PPDM asks NAS and asks if there are new shares
 - if this no name share, PPDM cannot scan for new shares, no one to ask for new resources to backup

29:00 PPDM supports any CIFS / NFS
 - huge fast backup for any share
 - for self-backup infrastructure we need supported system to ask about new resources

29:34 - Do we need any post-backup for always FULL?
- NO any post backup actions required
- We have FULL backups like that!
- Please test it and touch it on your own

30:15 Restore
 - we have many backups of different shares
 - backups of NetApp / Qnap
 - we can restore to the same NAS
 - we can restore to Linux / Windows
 - we can restore to any target
 - we do not use NDMP 

31:52 Granular restore
 - we can restore single file

32:15 We can search for files

32:38 Summary
 - Great solution that can help with time, many files, large volumes
 - Very cost effective
 - Look more closely
 - Future - show the numbers
 - I am very excited about technology
 - See you next time!

Presentation from the event:

Backup and Recovery - all materials:
BackupRecoveryGuy: Backup and Recovery materials
Fast backup of files - video:
Fast backup of files - article and presentation:
Performance in backup:
DD6400 video: 
DD6400 article: 
New generation of Data Domains DD6900 / DD9400 / DD9900
What is new in Data Domain 7.2?

Only successful restores!