Thanks for explaining those concepts so well. The only difference from this 2015's video and now is that you could use 3 fault domains and up to 20 update domains.
Aboslutely wonderful description. I am reading through the course on Microsofts website and "Fault Domain" was confusing me. Your explanation cleared everything up instantly. Thank you.
After watching so many videos this one is the only one that nailed it, such great info, loved it so much I had to login youtube just to leave that like
thanks for the explanation. when apps from 2 different availability set (AS-db, AS-web) with update domain 3 and fault domain 2. both have 5 vms. vm1 from both (vm1-app, vm1-db) goes to update domain 0. now Microsoft did maintenance on update domain (0). will both the vm's (vm1-app, vm1-db) will go for reboot? will it be applicable to all availability set. meaning vm1 (first VM update domain 0) from all availability sets under a Resource group will reboot when Microsoft do a maintenance on update domain 0
awsome explanation john...can u plzz explain types of disk in brief i.e OS, temp and data disk.Even difference between manage vs unmanage disk and what is the storage stamp
This tutorial explains concepts of Availability Set - Fault and Update Domains very well. Thank you very much! To clarify the part on update domain, you are making a reference of updating the web server hosts - IIS correct?
right update domains apply to the individual VMs which for IaaS would impact when the physical box hosting the VM has to be updated to a new image. For PaaS could be new version of app rolling out.
Thanks John. , let me test my knowledge on this topic . Let say i have 14 VMs in an Avset in 2 FD and 10 UD. what is maximum VMs Unavailable for each UD and FD? i calculated 2 VMs in UD and 7 in FD, correct?
The thing if we use availability set do we need add LB to makes sure VM is able to route to public seemlessly? If that so, do we need to make sure application or data is replicated across VM? Or its all done automatically and i don't need an external LB or make sure my content on both VM are sync.
You have to make sure data is sync'd across if its stateful and yes if want to make a service available across multiple need a single endpoint for clients which is load balancer/app gateway etc.
Hi Mr. John, I want to know if for eg, I place 2 web servers in 2 fault domain and 3 update domain. does that mean there will be in total 5 copies of VM created for HA? Also, when there is a request coming into my Web Server, will it be served by these 5 copies of VM or is it only the 2 Web Servers? Will my web servers placed in Fault domain and update domain act as Active Active or Active passive servers (i.e, update domain VM kicks in when Fault domain VM goes down). Please explain. I am not quite getting the concept.
Does not work that way, they are 2 different dimensions. If you have 2 web servers you can only be in 2 fault domains (blast radius) and 2 update domains (update rollout). 1 web server in fault domain 1 and update domain 1 and 1 web server in fault domain 2 and update domain 2 (for example)
If 2 VMs on a update domain are updated and they are running a new version of the app and other VMs running older version, the what version would be avaailable to the end customer?
Depends which instance the load balancer in front pointed them to but would be a mix. If you can’t have a mix, ie maybe the database schema changes as part of an update that they talk to, then you have to look at different deployment strategy like blue/green and flip between them
Thanks for explaining those concepts so well. The only difference from this 2015's video and now is that you could use 3 fault domains and up to 20 update domains.
Very well explained about fault domains and update domains , fault domains = unplanned maintenance and update domains = planned maintenance.
Aboslutely wonderful description. I am reading through the course on Microsofts website and "Fault Domain" was confusing me. Your explanation cleared everything up instantly. Thank you.
I was struggling to understand the fault and update domain. Thanks for explaining very well.
Yeah! Now Everything is clear to me too.
I found your session to be the best on explaining the topic in depth among the videos I tried to clear my concept...Appreciate your effort
Thanks. Glad it helped.
Big help!!!!!! Thank you from Philippines with love!!!!
Great Job! Couldn't get my head around the update domain part till I watched your video and the light bulb went!
After watching so many videos this one is the only one that nailed it, such great info, loved it so much I had to login youtube just to leave that like
Thank you!
Thanks John. That was very clear and perfect explanation. I was struggling with the difference in fault and update :D
Same here
MSLearn did not make this topic easy to understand. You did it perfectly, i dont know how you are so good
Awesome. Thank You John. It can't be explained better than this.
super explanation in between fault domain and update domain, thanks John
Ah, finally it all makes sense now, thank you, worth subscribe to the channel to watch more.
I finally understood how update and fault domains work. Thank you
Awesome explanation. Made it crystal clear . Thanks John.
This video manage to bludgeon this concept into my thick skull! 💀
Thanks a lot, John! Great explanation. You earned a sub!
Thanks
Yes! Great explanation. Thanks for creating this content.
Welcome
This really helped me to understand this topic, many thanks.
Thank you John.
here in 2024 . you are a star John
lol
thank you John, it's very clear explanation
Glad it was helpful!
This is a perfect explanation. Thank you.
John use to lift server racks. And bench press power units
this was really helpful... Thank you
Very welcome!
excellent. but pls use some kind or a wireless microphone so there wont be any echo.. Thank you.
Awesome explanation! Thank you so much.
Thank you fella
🤙
Very helpful and clear. Thank you!
Super explanation - thank you.
Thank you, very nicely explained..
Thanks
This is a great explanation! However, I didn't follow the last bit stating in IaaS we don't have this concept.
Thanks a lot John !!
fantastic video.
Perfect, thanks great explanation, subscribed your channel right now.
Awesome, thank you!
thanks for the explanation.
when apps from 2 different availability set (AS-db, AS-web) with update domain 3 and fault domain 2. both have 5 vms. vm1 from both (vm1-app, vm1-db) goes to update domain 0. now Microsoft did maintenance on update domain (0). will both the vm's (vm1-app, vm1-db) will go for reboot?
will it be applicable to all availability set. meaning vm1 (first VM update domain 0) from all availability sets under a Resource group will reboot when Microsoft do a maintenance on update domain 0
FD and UD are per availability set. Not across availability sets.
Thank you
awsome explanation john...can u plzz explain types of disk in brief i.e OS, temp and data disk.Even difference between manage vs unmanage disk and what is the storage stamp
amazing.. Keep it up..
Thank you.
Awesome! Thanks
Nice tutorial thanks. Just to confirm you said 5 fault domains at 8:50 did you mean Update domains? where PaaS has 20?
so 3 fault and 5 update by default but can go up to 20 :-)
Nice tutorial
This tutorial explains concepts of Availability Set - Fault and Update Domains very well. Thank you very much!
To clarify the part on update domain, you are making a reference of updating the web server hosts - IIS correct?
right update domains apply to the individual VMs which for IaaS would impact when the physical box hosting the VM has to be updated to a new image. For PaaS could be new version of app rolling out.
Thanks John. , let me test my knowledge on this topic .
Let say i have 14 VMs in an Avset in 2 FD and 10 UD. what is maximum VMs Unavailable for each UD and FD? i calculated 2 VMs in UD and 7 in FD, correct?
Yes but normally minimum 3 fds
The thing if we use availability set do we need add LB to makes sure VM is able to route to public seemlessly? If that so, do we need to make sure application or data is replicated across VM? Or its all done automatically and i don't need an external LB or make sure my content on both VM are sync.
You have to make sure data is sync'd across if its stateful and yes if want to make a service available across multiple need a single endpoint for clients which is load balancer/app gateway etc.
Would it make sense for the fault domain count to be the same as the update domain count, if availability is the main goal?
No, fd is limited to 3 while Ud can be higher which would limit impact during planned operations.
thanks..this helped
Hi Mr. John, I want to know if for eg, I place 2 web servers in 2 fault domain and 3 update domain. does that mean there will be in total 5 copies of VM created for HA?
Also, when there is a request coming into my Web Server, will it be served by these 5 copies of VM or is it only the 2 Web Servers?
Will my web servers placed in Fault domain and update domain act as Active Active or Active passive servers (i.e, update domain VM kicks in when Fault domain VM goes down).
Please explain. I am not quite getting the concept.
Does not work that way, they are 2 different dimensions. If you have 2 web servers you can only be in 2 fault domains (blast radius) and 2 update domains (update rollout). 1 web server in fault domain 1 and update domain 1 and 1 web server in fault domain 2 and update domain 2 (for example)
If 2 VMs on a update domain are updated and they are running a new version of the app and other VMs running older version, the what version would be avaailable to the end customer?
Depends which instance the load balancer in front pointed them to but would be a mix. If you can’t have a mix, ie maybe the database schema changes as part of an update that they talk to, then you have to look at different deployment strategy like blue/green and flip between them
Last 1 min of the video is the most important part
Update domain needed for maintenance and fault domain for min. Downtime.
Thank me later !
Rack city