cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
855
Views
0
Helpful
3
Replies

Hyperflex edge multisite

mikhailov.ivan
Level 1
Level 1

Hi all! I have couple question about Cisco HyperFlex Edge Architecture.

For example:

I have 3 site with 3 HF node HX220c and between site i have WAN channel (internet\sd-wan)

1)What requirment for WAN throuthput in this case? Minimum.

I'm afraid what HyperFlex sync will utilize all my wan channel. And how often they sync?

 

2)I can build 3 independent HF node (per site) or its one storage for all sites?

Thanks!

1 Accepted Solution

Accepted Solutions

Kirk J
Cisco Employee
Cisco Employee

What Hyperflex 'sync' data are you referring to?  Hyperflex guestVM replication/DP traffic? If so, min hx replication traffic requirements are:

  • "The allowed low bandwidth is 10Mb and the maximum latency supported with 10Mb is 75ms. If the initial replication of VMs fails due to lossy network or unstable HX clusters, the VM replication will be initiated again in the next schedule as a fresh replication job. In this case, you have to size the schedule accordingly to protect VMs."

As to question2, if you are asking about stretch cluster, https://www.cisco.com/c/en/us/td/docs/hyperconverged_systems/HyperFlex_HX_DataPlatformSoftware/HyperFlex_Stretched_Cluster/3_5/b_HyperFlex_Systems_Stretched_Cluster_Guide_3_5/b_HyperFlex_Systems_Stretched_Cluster_Guide_3_5_chapter_010.html, that has pretty stringent bandwidth requirements of 10GB:

  • 10 Gbps dedicated, 5-ms Round-Trip Time (RTT) latency between the two active sites is required.

    100 Mbps, 200 ms-RTT worst case latency for 16 kilobyte packet sizes between the active sites and witness site is required.

 

 

Kirk...

View solution in original post

3 Replies 3

Kirk J
Cisco Employee
Cisco Employee

What Hyperflex 'sync' data are you referring to?  Hyperflex guestVM replication/DP traffic? If so, min hx replication traffic requirements are:

  • "The allowed low bandwidth is 10Mb and the maximum latency supported with 10Mb is 75ms. If the initial replication of VMs fails due to lossy network or unstable HX clusters, the VM replication will be initiated again in the next schedule as a fresh replication job. In this case, you have to size the schedule accordingly to protect VMs."

As to question2, if you are asking about stretch cluster, https://www.cisco.com/c/en/us/td/docs/hyperconverged_systems/HyperFlex_HX_DataPlatformSoftware/HyperFlex_Stretched_Cluster/3_5/b_HyperFlex_Systems_Stretched_Cluster_Guide_3_5/b_HyperFlex_Systems_Stretched_Cluster_Guide_3_5_chapter_010.html, that has pretty stringent bandwidth requirements of 10GB:

  • 10 Gbps dedicated, 5-ms Round-Trip Time (RTT) latency between the two active sites is required.

    100 Mbps, 200 ms-RTT worst case latency for 16 kilobyte packet sizes between the active sites and witness site is required.

 

 

Kirk...

Thanks Kirk!

1) Yes thats wthat i mean! but I couldn't find answer on Cisco.com (

10mb it look like a true.

 

2)I mean for example, if i have 1 VM on some node on one site and all 3 nodes on this side is die simultaneously a VM still work (or restart) on another site? Or VM replication only in one site nodes?

Thanks for help!

So, you can have multiple replication groups, that go in both directions (i.e. site 1 has a,b,c VMs replicating to site 2.  Site 2 has VMs e,f,g replicating to site 1)

Please take a look at https://www.cisco.com/c/en/us/td/docs/hyperconverged_systems/HyperFlex_HX_DataPlatformSoftware/AdminGuide/4_0/b_HyperFlexSystems_AdministrationGuide_4_0/m_hxdp_replication_vm_disaster_recovery_4_0.html

 

Kirk....