Rubrik, Cohesity, Veeam, curious if anyone has any experience with them? PoC experience to share, other thoughts?
Thanks Adnan, i appreciate the feedback.
We wanted to remove all the unnecessary layers of complexity (backup software, backup servers, backup proxies, catalog DB, backup storage / archive, etc.) and replace it with a hyper converged appliance. Beyond simply the software as a solution, we were interested in the appliance aspects of the overall global solution.
We also wanted condense, simplify the backup software and globally dedupe storage into a single, scale-out fabric, being able to easily grow as you go without limits, just like other hyper converged platforms such as Nutanix or Simplicity.
Having the ability to archive out to the platform of our choice was important, whether that's on-site / off-site object store, or even the public cloud such as Amazon - AWS, Microsoft - Azure or Google - GCP.
There are number of new updates to the Veeam backup and replication product, unfortunately they won't be GA until late this year or early 2019 and i'm not interested in seeing beta features for a PoC. We also had some concerns with the lack of global deduplication, its per job schedule was my understanding.
Hi Michael, understood your point. We don't have experience with an all-in-one solution. Yes Veeam Dedupe is per job basis, we used Dell DR4100 deduplication appliances for storing the backups. Both Veeam and Dell DR-4100s are very simple to setup, minimal management and work great.
Content you might like
Slow recovery response times35%
Data availability is limited47%
Too expensive to scale effectively52%
Difficult to manage for widespread use38%
Prone to misconfiguration12%
No - There are no drawbacks7%
Using CSI plug-ins to support containers with existing infrastructure18%
Hosting all container apps in public cloud using native cloud storage56%
Using container native storage such as Portworx, MayaData, OpenEBS, Rancher7%
Have not figured it out18%
Other (please comment below)1%
Thanks, we are getting ready to run a Poc for Cohesity and Rubrik. Can you expand a bit on what the Rubrik point of failure is,
Here are some additional details around the SPOF
If a node fails or is taken down for maintenance, any backup running on that node fails and has to be restarted. This could mean missed SLAs which may be unacceptable for most customers especially since everyone has 24x7x365 operations.
The exception is Oracle RMAN backups which are sent to the node that has a floating IP address. Since the floating IP address can fail over to another node, Oracle RMAN backups don't fail on node failure.
The reason behind this is that a Rubrik cluster has only one floating IP which lives only on one node. This floating IP is used for what they call managed volumes. These managed volumes are used only for instant VM recoveries and Oracle RMAN backups.
https://www.cohesity.com/blog/cohesity-pegasus-6-1-simplifies-management-secondary-data