Search This Blog

Friday, October 30, 2015

Day 4 - VeeamOn 2015

We had an awesome Veeam Party yesterday. Had a difficulty to woke up and now I need a cup of coffee so I can attend today last session.



Here is our note taking during today session.

Session 1: Veeam Backup and Replication Worst Practices (tom)

Veeam - It's Just Work! (in the logo)

 

But we

  • Install veeam with default

 

Bad-Ignore the infrastructure

  • Enough space on the datastore
  • Misses pDRMs or in-guest iSCSCI
  • Fails to properly size repository size
  • No monitoring for critical issues (snapshot, latency)
  • Store backup on VMFS can complicate recovery
  • Veeam components may not be sized for environment and especially growth
  • No off site recovery options from catastrophic disaster.

 

How to Avoid?

Use Veeam ONE

Before Install

  • VM configuration assessment
  • VM change rate estimation

After Install

  • Monitor capacity (repo space, datastore)
  • Early warning (snapshot)

 

Don’t Ignore the infrastructure

  • VCEnter or SCVMM performance is critical
  • Pay close to DB maintenance plan
  • Don’t reboot vcenter when backup are running

 

Consider all external loads

#1 cause of random backup failures is external infrastructure load

 




One Job Guy

  • Configure all VM into one backup job!

 

One Job Per VM Guy

  • Want a complete control of policy per vm

Bad! -

 


The Job Chain Guy

  • Guy that always job…job…and chain it.
  • Manually chain jobs rather than using the automatic scheduler
  • He wants to control the load


 

The CheapSkate!
 

 
 
The Consolidator


 
 
The Tweaker
 


 
That's all for different types people that can make your deployment turn to worse. If you do have any feedback , please feel free to use Veeam Forum


That's all for today sharing folk.

Besides than attending general session and closing ceremony, we had an opportunities to join ViaWest DataCenter Tour. It was an awesome site visit. Too bad! we are not allowed to take any picture during tour. Probably next year you can reserve a seat to join on next year VeeamOn.

See you folk next time and have a safe journey home!

Day 3- VeeamOn 2015

Day 3 - we started with a general session which talked about how Veeam work closely with ecosystem to ensure Always On! and Availability.

Ecosystem:
  • Microsoft
  • VMware
  • HP
  • Cisco
  • NetApp

Veeam with HP


Veeam with Microsoft

New Update - Microsoft has 24 regions datacenter for Microsoft Azure.
Azure Software load balancer will be available on premise. Coming soon!


 
 


Session 1:- Scale Out Backup Repository (Luca, Mikhail) - SOBR

Problem:- Repository have a fixed size. A ton of wasted space

Add multiple repository to leverage multiple thread and create one job per repository

Scale out repository - group multiple simple repository into single entity

No nesting of scale out backup repository . Only 1Scale Out Backup Repository

Data locality
- Put all dependent backup files are placed on the same extend.

Performance
- All independent incremental backup files are placed on a different extent from the corresponding full backup file.

Previous version:-

]
 
SOBR:-

 
Full backup (.vbk) on one repository
Incremental backup on one repository


SOBR Architecture:-
When a job is started, select the best extent from each group by
a) Skip all unavailable extent
b) Skill all extent with no slot left
c) Select the extend with the most estimated free space available

Data locality - per VM

How to migrate from existing repository?
- Import existing repository to SOBR
Backup set will remain in place - no moving. Correct placing happen on the first run of the job after configure SOBR
Existing backup job can be remapped.

Unsupported - replica metadata, configuration backup and VM copy -> cannot use SOBR

Maintenance ( Yes - it is available for SOBR)

Evacuation when no longer require.

Manual file movement
- you can manually move files across extents
- After relocation, just run RESCAN

Final warning:-
Be careful with deduplication devices (they do synthetic operation internally)

If got dedup storage - don't turn on performance policy.

SOBR - will not support Cloud Connect Backup in V9

Session 2: Secure and Protect the Microsoft Cloud Platform with Veeam, StarWind & 5Nine Software (Symon)

5Nine Cloud Security - www.5nine.com/CloudSecurity



Feel free to check our previous blog about 5Nine Security Solution:-
5Nine uses it proprietary Change Block Tracking (CBT) driver to scan only changed block on the disk. Scan up to 70x faster.

Automate Security Task Management
-Powershell support
-Task scheduling

5Nine Cloud Security Management
-5Nine Console
-5Nine Powershell
-Azure Pack Extension  (FREE add-on for 5Nine Cloud Security). WAP allow you to run Azure services in your datacenter on your hardware.
-SCVMM with 5Nine Plugin (FREE add-on for 5Nine Cloud Security)


Next product:- 5Nine V2V Easy Convertor (FREE)
-VMware to Hyper-V Conversion

Virtualization Management - 5Nine Manager
- included centralized dashboard
-Support Hyper-V Replica management. VMM do not support management for Hyper-V Replica.
-Monitoring built-in
-Inventory Report
-Best Practice Analyzer
-Included Logging. Easy to view log
-Include Library similar like VMM Library to store template & sysprep VM option is available
-5Nine Manager with Antivirus. Agentless Protection. (without IDS & Firewall)

StarWind Products:-
  • StarWind Virtual SAN
  • StarWind Hyper-Converged Platform
  • StarWind Virtual Tape Library (VTL)

Something new that we found out that you can manage StarWind Virtual SAN & Veeam Backup and Replication by using 5Nine Console. Integration in-place between Veeam, StarWind and 5Nine.



Session 3: Top Investment in the Next Release of Hyper-V (Ben)










 
VM Storage Resilient
when storage has problem, VM will paused. Nothing write in disk. It will resume when storage is up and running without administrator intervention.
 

 
VM Compute Resiliency
When network outage occur, wait for a short period of time before failover. default is 4 min (can tune). If 2 min, we will wait.
 

 
Quarantine of flapping nodes
When a compute got intermitent issue, the problem come & go. Node leave the cluster and back later.
1. Keep track how many time it left and join. Default -3 times an hour.
2.If left more than 3 min, it will prevent node from joining to cluster and no vm placement. Default action is manual intervention to remove node from quarantine.

 
New backup infra
- no vss in host (still available on guest)

 
Shared vhdx
  • VHDX resize without downtime
  • guest cluster with shared vhdx can be protected by Hyper-V Replica
  • agentless backup of shared VHDX (guest cluster)
 


Online VM Configuration Changes
  • network adapter can be added/remove while Gen 2 VM running
  • resize static memory when VM is running.


Cluster OS Rolling Upgrades
  • rolling update with Win 2012 R2 and Win 2016 nodes within the same cluster
  • zero downtime cloud upgrades for Hyper-V and Scale Out File Server
* Maintain functional level and perform testing before change functional level

 
Operational Improvement
- Production Checkpoint. (By default). We are supported in production environment starting from Win 2016. It is safety net. Instead of capturing memory we do VSS snapshot to ensure consistency

 
Powershell Direct
- no dependency on network. Can access and perform action to guest VM from the host.

 
ReFS Accelerated VHDX operation.
Key word is "quick". Not instant! :)







RemoteFX client on latest Windows 10 Build. Try it out!
Nested Virtualization - latest Windows 10 Build- first OS. More to come ...

Demo Video 1 - https://youtu.be/G2p5sEs-XeY
Demo Video 2 - https://youtu.be/OrjKUTqb8jo
Demo Video 3 -https://youtu.be/bfrPZvFjW4k

Veeam Party Night
After a whole day sessions, finally we all went to Omnia NightClub at Caesar Palace. Here are some of the picture during Veeam Party.







Well , we enjoyed the party. They are a saying "What happen in Vegas, Stay in Vegas"..hahaha...we didn't do anything bad. Just enjoyed the party.

Alright that all for Day 3. Stay Tuned for Day 4...