- Is this per site or possible to have distributive setup which installed on different site?
-"We only allow a single APIC in lab/test environments where redundancy is not required." So meaning you can still use a single server to control the devices?
But in this statement
"When the connection between two sites is lost, the site with one APIC controller will be in the minority (site 2 in the figure above). When a controller is in the minority, it cannot be the leader for any shards. This limits the controller in site 2 to read only operations; administrators cannot make any configuration changes through the controller in site 2"
So meaning I cannot use a singe APIC server to control the infrastructure?
2. Split brain condition?
- What does split brain condition means? Is it for 2 APIC server or This is for multiple sites with distributed APICS?
- Can give simple scenario that split brain occur?
Re: Increasing APIC Size and Split brain condition?
For production networks, a minimum of 3 APICs is supported and anything less is not. Anyone who wants purchase an ACI lab kit (IIRC is 1 APIC, 2 leafs, and 2 spines), then 1 APIC will work for lab/testing environments. There is configuration in the APIC which will permit a single APIC to be a single cluster.
In regards to your first question, the minimum is 3 APICs per fabric. For stretch fabric or multi-pod, these are still considered 1 ACI fabric even though they may be in different sites. For multi-site, each site is considered an ACI fabric.
With ACI, the APIC clustering is done in odd number of APIC servers (1, 3, 5, etc) to allow a majority or minority state.
Lets say you have a target cluster size of 5 APICs. At some point a failure occurs and now APICs 4 and 5 are unable to communicate with APICs 1-3. APICs 4 and 5 are able reach each other and form a healthy cluster. APICs 1-3 are able to cluster together without APICs 4 and 5. You're now in a situation where there are 2 different APIC clusters are in conflict and the switches need to know which APIC cluster should they receive any new policy from. This is where the split brain condition comes in. APICs 1-3 would be considered in majority state (3 out of 5) whereas APICs 4-5 are in minority state (2 out of 5). In any case where APICs are in minority state, they cannot create new policies and make changes. Anytime there is a majority cluster, then changes are safe to make.
How does everyone feel about service graphs in a brownfield network centric cutover? maybe I'm an old school engineer but I was taught that if you have to do a Pbr then your design is wrong. they are a pain to troubleshoot because you always f...
To participate in this event, please use the button to ask your questions
This special event - formerly known as Ask the Expert- is open only to Cisco Customers and Partners.
Many pages in the Cisco Community are acce...
Going to Cisco LIVE EMEAR 2020? Give your Data Center agenda a boost with these top recommended sessions. Register and join us!
Data Center Networking:
From Reactive to Proactive NetOps With Cisco Network Assurance and Insights- The Power of Data...
Hi allWe're trying to tackle a DHCP relay problem between a set of N9K-C9332PQ (NXOS 9.3.2) and dnsmasq (doesn't work) resp Windows 2016 DHCP service (works with some tweaking).SITUATION:We have a BGP EVPN VXLAN (L2VNI and L3VNI) with anycast gateway setu...