NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
m4300 m5300 mlag support
Are there any plans to support MLAG on the M4300 line of switches? We recently moved from m5300-52g3 to m4300-24x24f and while we are using stacking, support for mlag would be very useful. The o...
Hi mstinaff
Thank you for your message. There is no ongoing MLAG development for our M4300 line, instead we are working hard on other enhancements for M4300 (MLAG isn't ranked high amongst customer requests).
To LAG across several stacks certainly represents a challenge beyond <<STP at Layer 2 and VRRP at Layer 3>> limitations. At this stage, we can envision another tier, just above M4300 stacks, that would provide distributed LAG to your hosts, and distributed LAG to our M4300 stacks. But I understand this may modify your network topology.
Regards,
Thanks for the reply, just wish it was better news.
We are trying to avoid any single point of failure so if I understand your suggestion we would have another tier of redundant 10g switches to provide mlag to the servers and existing m4300 stacks which will be able to connect as though it is a normal single switch lacp lag, yes?
I'm not eager to justify the expense for another set of redundant 10g switches, Any other suggestions for approximating mlag functionality with a pair of m4300 stacks,
-a pair of m4300 10G stacks
-a m5300 L3 stack that was replaced by the m4300s (to get 10g)
-one xs712t switch that is currently serving as an aggregation switch of sorts
-a stack of m4300 1g switches
Hi mstinaff
Based on your topology, yes I think the best move would be a new tier above all your existing stacks.
I would consider this as an aggregation upper layer, and use a pair of M4300 10G switches, stacked together and dedicated to it.
With Seamless Failover, this pair of 10G switches would never stop if all downlinks to your existing stacks are distributed across the two switches. This is distributed link aggregation, not MLAG. For these LAGs you should use LACP default hashing, or if you setup the same on your other stacks, more advanced hashing algorithm based on IP and TCP/UDP ports.
Last, don't forget to also use Distributed LAG on the other side (on your existing stacks) with the master switch, and the backup master switch each time.
With this mesh, all based on LACP, you would achieve real, unstoppable 10G aggregation.
I hope this helps!
Regards,
Based on your topology, yes I think the best move would be a new tier above all your existing stacks.
I would consider this as an aggregation upper layer, and use a pair of M4300 10G switches, stacked together and dedicated to it.
With Seamless Failover, this pair of 10G switches would never stop if all downlinks to your existing stacks are distributed across the two switches. This is distributed link aggregation, not MLAG. For these LAGs you should use LACP default hashing, or if you setup the same on your other stacks, more advanced hashing algorithm based on IP and TCP/UDP ports.
Last, don't forget to also use Distributed LAG on the other side (on your existing stacks) with the master switch, and the backup master switch each time.
With this mesh, all based on LACP, you would achieve real, unstoppable 10G aggregation.
I hope this helps!
Regards,
Hi Laurent,
Distributed link aggregation is great and we've been using it from day one. Previously had four m4300 10g in a single stack with critical servers using quad sfp+ to lag across all four. But firmware updates or anything that required a stack reboot meant scheduling downtime to accommodate the short but inevitable network drop.
We have since then reconfigured them as two stacks of two so if needed the stacks can be restarted separately with no network loss. We've got bridging solutions set up on the servers but it is not as clean a solution as mlag would be (at least from the server configuration perspective.)
For 10g going forward, which products will have mlag to mitigate stack reboot outages?
Or is it possible to get a supported rolling update/reboot on live stacks? Please?
I am pleased you find the M4300 Distributed Link Aggregation feature great. Hosts, servers or any network device can be multi-homed to several switches in a stack for added hroughput and resiliency.
I confirm we don't show roadmap for MLAG, beyond what we previously achieved with our M6100 and M7100. We believe <<Non Stop Forwarding>> stacking is the way to go instead, for our top of the line, fully managed stackable switch series.
Yes, we hear your voice and consider rolling firmware update on a stack without reboot should be a strong, valuable enhancement. I am adding your vote in our list of customer enhancement requests. I'll be happy to share about this, when possible. Other contributors in the Community, please add your voice here!
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
