Hi Everyone,
I was hoping to get some general guidance on choosing a percentage for admission control. My situation is pretty basic. I have three identical nodes running ESXi 5.5 in a cluster. I'm trying to ensure that if one of the nodes goes down, all the down servers will be able to restart and for the most part, have similar performance as before.
So in enabling admission control, I was thinking I would reserve 33% of my resources for HA. I can check my VM's for memory reserves and increase my reserve percent accordingly. But as a general rule, if I have 3 nodes and 33% reserved and a node dies, will I generally get similar performance as before? Any gotchas or additional considerations I should account for?
Thank you in advance!
Welcome to the community Mike
I always preferred the “Percentage Based Admission Control” policy. the performance will be same.Also many depend upon the reservation set for vms and the workload during HA failover
The best you can get it from Duncan Blog here. I learnt a lot from this and i believe you will too. Percentage Based Admission Control gives lower VM restart guarantee?
Good Luck
Welcome to the community Mike
I always preferred the “Percentage Based Admission Control” policy. the performance will be same.Also many depend upon the reservation set for vms and the workload during HA failover
The best you can get it from Duncan Blog here. I learnt a lot from this and i believe you will too. Percentage Based Admission Control gives lower VM restart guarantee?
Good Luck
After building your HA policy you can test it in HA simulator. HA Simulator : What will happen to your VM. | Vmwareminds
Hi Anjani,
Thanks for sharing the info about HA simulator. Seems like amazing tool as i gone through your blog. I will share my experience after testing this in my LAB and if can get permission to simulate this in our production. Thanks once again.
Your always welcome . you need to just pull out the DRMDUMP from your vcenter. and upload that dump in to this website. it will show you if your settings are good enough to handle a failover.
I dont think so that management mind it as it will never impact your running environment. you can say you are taking precaution and fixing any issue proactively instead waiting for a complete shutdown and respond Reactively.;)
I really appreciate the help. Thank you!