28. #VMware Administration v6.7 -Distributed Resource Scheduler (#vspheredrs)- Part 2

so what we will do we will test the drs nav just to just to reiterate the current lab setup what we have okay i have three hosts one two three these three and i have around nine virtual machines nine virtual machines and three data stores shared data stores i'm talking about sand storage so go to cluster and data stores you have three local forget about them and three sam okay three sand data stores now i want to identify which all vms that are running on local and that are running on sand okay how to do that go to storage tab and vm1 is running on local storage one let's get that into san storage okay we need to bring that to sand storage that is first thing second thing six and third third one seven these are the three vms which are running in my local storage okay so the what are the prerequisites just to reiterate once again the prerequisites are for drs testing if drs on hc rules plus vmotion kernel port okay agree now tell me the h rules what are the rules that are you need vc right you need to create a cluster and shared storage finally proper network right these are the things that we need mandatory now now h is already tested means all all these rules are satisfied and vmotion kernel port we have enabled yesterday yesterday we were unable to test it because the storage is not visible even today also that we got the same problem that is the reason why i have rebooted the entire lab again now the lab is up and running we can do the testing so v motion is already enabled let's check to make sure okay go to configure kernel adapters one v motion is there and four other kernel ports on this host let's see a lot of lot of services are there but vmotion is enabled here the same on each host it's same on each host motion is enabled and the ip is 23 223 222 and 21 fine okay now now my question is do we have all these nine machines or on a shared storage or on a local storage okay i'll put it on red put it on different color yeah i have one local storage here i have one vm in local on this host also one vm in local on this host also one vm is running on local on this host vm1 on this host vm6 on this host vm 7 are running in my local storage that is what we identified right so now virtual machines are up and running without any downtime i want to migrate this vm from this storage to the storage understood okay if i want to migrate virtual machine from one storage to another storage you call it as sv motion storage vmotion how to do that let's see quickly select the virtual machine which you want to migrate it is now in local look at it is now in local right click migrate storage only okay select storage only you want to change the storage not the host next select the storage where you want to migrate i'll migrate it on to learn 20 fine finish go to recent task see machine is relocating and it is now migrated to 120 without any downtime if you want this machine doesn't have any ips and let me go to second host okay let me go to vm6 vm6 is now running on local right okay let's take the console let's take the remote console and i will assign the ip do the sv motion it should work without any doubt control panel network what is the vlan what is the vlan for these vms 192 168 summary will and 40 okay 40 dot 40 okay and one okay now come out and minimize this ping 192 168 40.40 it's working right the ping is continuously working this machine now this machine is running on with storage vm6 is running on local storage now let me migrate storage only i'll put it on let's see store machine is moving but you see no downtime nothing machine is already relocated from one storage to another storage okay now you see machine is running on sand storage which is 130 but server is up and running there is no interruption understood so migrating machine from one host to another host what you call it as quickly sorry motion migrating virtual machine from one data store to another data store it's view motion okay is there any downtime in both the cases no no okay now now there is another option migrate both okay migrate both the host where you want to migrate vm6 next select the data store so it is migrating both virtual machine and virtual machine from host to host and storage to storage both at a time okay you see there is a little downtime earlier it this option is not not there and 6.0 5.5 what i have worked up on you need to bring the virtual machine down if you want to do the both the things at a time please be on note okay if you want to migrate virtual machine from one host to another host at the same time one store to another storage at the same time at the same time you want to do the task then you need a downtime but in 6.7 that is eliminated a fraction of ping time out then you can relocate okay everything is everything is online there is no down time so vmotion majorly majorly vmotion will be used to avoid downtime and moving resources here and there within the cluster or within the within the uh storage within the host there are multiple data stores are there so you can migrate here and there right now now let's go back to now let's go back to our discussion is on drs now vmotion is up and running h is configured properly all the options are satisfied and i can even tested migrating vms from one data store to another store and migrating vms from one host to another host both i have tested okay vmware drs is all about moving virtual machine automatically from one host to another host one host to another host that's it not the storage okay if you want to automatically load balance the storage then you have to configure sdrs storage distributor resource scheduling or resource scheduler that we will do it tomorrow in in this session what we will do is we will test how drs will work fine any questions after this okay so what is drs i said simple feature where it will optimize your resource utilization within the within the cluster means if this horse is ninety percent used this is ten percent use this is fifty percent used it will automatically migrate one small vm or a big vm based on the recommendation from here to here so that so that this this host utilization will increase to 50 percent and this host will utilization will decrease from 90 to 60 percent so this is running on 50 this is running on 50 this is running on 16 so everything is satisfied now you have load balanced all the resources across cluster that task can be achieved by drs understood you getting any any background noise or something not now now yes i'm getting some yeah just a moment i'll close the door i'm sorry yeah now clear yes okay so drs drs will help you to load balance the resources in multiple ways okay in multiple ways basically there are three ways to do the there are three ways to do the load balancing what to do okay still it is speaking it's okay okay fully automated drs partially automated drs manual dns there are three options okay let's understand when this drs will come into picture now imagine i have nine vms when the drs will run i said high availability is 24 by 7 it will send a heartbeat to master in high availability heartbeat will be heartbeat will be send it to master 24x7 if the heartbeat is missing for a specific duration then master will come into picture and do the restart that is what we discussed in hca now in drs when the drs will come into picture and when it will work drs will come into picture in two situations okay in two situations one is one is when you power on any pm first option is when you power on any virtual machine within the cluster there is a virtual machine which is powered off currently the moment you power on drs will run okay otherwise otherwise the second situation is every five minutes every five minutes it will do the calculation okay we need to test both the situations understood hello yes okay now what i will do what i will do fully interpreted drs fully automated drs how to set go to your vcenter go to cluster go to cluster let me minimize this okay yeah go to cluster configure drs drs if you look at edit settings dis is enabled automation level fully automated partially automated manual drs that is what i have written there the irs okay and migration threshold you want to run the drs more aggressively whenever it will absorb small change small difference it will start migrating or liberal or conservative leisurely lazy you want to run more proactively at least repetitively it's up to you what i will do i'll put some aggressive because i'm running a small small vms calculation is bit difficult so if i say more aggressive there is a little change in the performance a little change in the resource utilization it is trying to calculate and do the migration understood okay yes done now it is set to fully automated drs when the drs will come into picture whenever you power on any virtual machine or on your power on any virtual machine how simply we can test this let me do one thing let me go to cluster be on mute be on mute let me go to cluster and i will power off all the vms okay i will power off all the vms let me add a couple of columns here post on which host it is running that's it at least i need this right okay now sort by host three four vms running on four vms running on host one two vms running on host two three vms running on running on host three now let's understand all the vms are running no no no see we forget one thing one one machine is running on local storage let me migrate that to sand storage you see otherwise otherwise it's difficult that machine will not come into picture that machine will not come into picture for calculation i'm saying right so let's go to cluster now all the machines look at the machines three three on one host and two on other host four on four machines on another host what is the first condition first condition is sorry what is the first situation when when you power on any virtual machine let me select all the machines all the machine and migrate its machines are machines are powered off now currently machines are powered off this is not the vmotion just you are re-registering the vms from one host to another host when the machines are powered off you don't call that as a vmotion just simply change the compute resources and i want to put this so zero two i want to put this one zero two next okay put the vms in respective venas it's okay fine finish if you look at if you look at it is relocating re-registering all the vms fine so now you see all the hosts are now showing us zero two means all the nine vms all the nine vms running on all the nine vms running on those two actually not running they registered on host2 okay now if you power on what happens drs will run and it will do automatic load balancing fully in fully automated load balancing will be done automatically when you power on if i right click power on you see machines will be restarted in different different hosts refresh and all the machines are rebooted now you see sort by host in one host three machines another host three machines another host three machines randomly you don't have any control they will automatically reboot or they will automatically not reboot power on in in load balanced way you see the load balancing is done automatically you did anything nothing simply enable fully automated drs power on machine will decide where to go and start understood this yes okay now this is in fully automated everything is a means automatic m means manual in fully automated situation okay when when you power on virtual machine machines will be load balanced automatically agree on this point that is what we tested now right now there is no new vms you created a virtual machine last to last week you don't have any new vms every virtual machine which is running on the cluster they are already powered on hold on long back few machines one year back few missions yesterday few minutes few missions day before yesterday doesn't matter okay all the machines are powered on now still in every five minutes drs will run every five minutes drs will run right so how to test that simply you have you have all the vms running on a different different host so let me select all of them and again migrate this time it's a v motion not the re-registering if a machine is powered off you are moving here and there then that is called registering the vm okay if a machine is powered on you are moving the vms from one hose to another hose here and there then that is called as v motion now i am doing a manual v motion select i will change compute resources now i will put everything into one host host01 i will put all the machines in host01 then see machines are now migrating everything will be on zero one i want to see everything on zero one see each and every machine now everything is on zero one right host 2 and host 3 both of them are under utilized host 1 is over utilized wait for 5 minutes drs will run and that will automatically move the machines that is what we expected right all the machines are up and running but one host is overloaded another host is not up to the mark or not not at all utilizing sitting ideal other two hosts are sitting ideal now drs will run it can calculate the utilizations and it will automatically move the machines let's see now when you relocate the machine it is showing as vsphere dot local administrator has done this task okay now what i will do if i want i can wait for five minutes then it will automatically migrate the vms system will automatically migrate the vms or else you can run you can initiate manual drs at any point of time how to go to monitor go to cluster and monitor monitor vsphere drs recommendations okay if you go to recommend vsphere drs and recommendation run drs now just click on it and watch this refresh recommendations and go to vms see mission 6 is now migrating the rest of the machines are in queue see who is doing initiatory system you are not doing anything system is doing for you okay you are not doing anything system is doing for you so it it migrated couple of machines let's see just refresh now you see sort four machines running on one host three machines running on another host two machines running on the host based on the recommendation it it did it in the sense vrs calculated the current utilization and drs decided that one host is running with nine vms another two hosts are running with zero vms so let me move three vms from host one to host two and two vms from host one to host three to load balance so everything is done by system itself what i did i should i don't want to wait for five minutes for the simple test so i came here and run a drs so it will calculate and do the stuff automatically otherwise otherwise even if you don't click on here it will run every five minutes clear on this concept or you have any questions yes no doubts this one is partially automated fully automated it's fully automated boss second condition it is also automatic during fully automated drs whenever you power on the machine whenever you pour on the machine it will check and load balance by themselves okay and also every five minutes drs will run and and it will do the load balancing automatically you don't need to worry system will check and do the load balancing means system will check and if it is if it is finding any uh resource utilization mismatch host 1 is high utilized host 2 is low utilized and host 3 is low utilized then it will automatically migrate i cannot wait for five minutes that is the reason why i have run it manually otherwise you can you can simply sit and watch it it will automatically do it after every five minutes understood yes okay in fully automated drs both the operations in both the situations system will do everything automatic now now let's me sorry let me let me go back and set this to partially automated go to cluster configure drs edit set to partially automated okay that's it you haven't changed anything else reconfigure the cluster partially automated drs in partially automated drs in partially automated drs let's see first condition first situation first situation power of all the vms again okay just power of all the vms and move all the vms onto host three okay now all the vms are on host3 why i am doing this manually because i want to simulate the lab real time you don't need to do all these things just create a vms system will take care of everything right just for our testing purpose i'm just recreating the examples now you imagine you have some virtual machines everything is registered on host three now you simply power on simply power on in partial automation okay but in partially automated drs now you see 3 3 three load balanced again understood okay in partially automated again when you power on when you power on virtual machine it is automatically doing the load balance task what is the difference between these two then no there is no difference it is as it is fully automated or partially automated doesn't matter when you power on the machine system will take care everything okay now let's test the every five minutes so i'll put everything live machines onto zero three again back to zero three okay now i'll put everything under 0 3 now it is migrating the vms anyways if you look at just refresh if you look at all the all the virtual machines are now running on running they are not found of their power on okay state power on on host three now now you have to wait for five minutes and see or if you don't want to wait for five minutes again go to cluster monitor drs recommendations run drs now okay if you run drs okay let's see what happens even if you run a drs or even after five minutes system will not do anything okay even if you wait for some more time it will not do anything what you have to do you have to go back to monitor once again and drs recommendations you see some of the recommendations are popped up here system is doing some calculation on the back end and it is recommending you saying boss i have done this calculation currently currently all the machines are running on host three so just to balance the average cpu utilization on host three with respect to other two host i want to migrate i want to migrate these machines on a priority means this machine to host to this machine host one this machine holds true this machine host one like this i want to migrate this six vms you want to migrate yes or no then you have to manually click on it apply recommendations then only it will go and do the migration preparing virtual machine for live migrating source host then like this it will migrate if you go back to virtual machines tab and see three three three it done the load balancing but but it will not do automatically you have to you have to specify here apply recommendations means means it will do calculation at every five minutes and it will show the result here whenever you feel like yeah today let's see what is the utilization okay so some of the recommendations are showing here in the morning you come and check if it is a smaller vm if it is a big vm or if it is a genuinely there is some issue and drs is not not running and you can run manual drs here okay in second situation during the partial automated whenever you power on system will do automatic load balancing and during the every five minute situation it is like manual you understood this concept the irs will run every five minutes drs will run but it will not take any action unless until you go and click it every five minutes it will calculate every five minutes it will calculate and show the results over there okay but it will not do any migration unless you approve it done last scenario let's test the manual drs how to test this go to placenter go to cluster configure drs automation manual drs okay set to manual drs now go to virtual machines and select all the machines and power off select all the machines and power off everything is powered off now move every machine onto one host i'll put it everything in host one see everything is on host one so what i will do i will power on first situation i will power on it will give some pop-up saying these are the power on recommendations okay place vm2 on host 1 vm 1 on host 3 4 on 2 1 1 3 1 2 3 like this it is doing all the load balancing for all the machines okay you have to follow this yes when you click ok then it will pour on when you click cancel it will not pour on okay right power on again it will give a recommendation when i say ok then only it will go ahead and load balance it means in manual automation sorry manual dns whenever you power on the task is manual you have to approve it otherwise it will not do the load balancing okay now you see everything is on load balanced condition three three and three let me migrate once again i'll put everything onto three okay relocate vm by so-and-so user let's wait see now everything is registered on 0 3 and all the vms are up and running on host 3 okay now go to monitor drs recommendations you can wait for 5 minutes or you can click and wait here once you click and wait it will automatically give you the recommendations once again you see was these are the recommendations you have to you have to take an action because one host is fully utilized another two hosts are sitting ideal so i want to do this to load balance are you going to approve it yes appropriate apply recommendations then it will go and migrate it will go and migrate the machines and if you go to virtual machines you see it got load balanced again on host number three four vms are running on host number two two vms are running and host number one three vm surrounding this is how it will do the load balancing this is manual now you understood what is the difference between three situations and two situations and three different uh automation levels yes okay so how how it is how it is migrating you have a kernel port here it is connected to one switch you have another switch here with two kernel ports another cable and you have another switch here with two second kernel port which will go and all these are connected in one switch so this vm if this vm want to migrate from here to here it will communicate via this kernel port not a vm traffic host will communicate and it will register here then this vm will come here this is how it is working on the backend okay why we need a drs and how it will work you understood on the prerequisites what we have done yesterday yes clear any questions up to this on the drs testing vivek is there or he left hello i'm there yeah anyway you are not going to do any of these tasks automatically okay sorry any of these tasks you will not you are not going to do it on a day-to-day basis system will do it automatically so if in our environment if in our environment we have some critical uh vms then we can enable drs uh in a manual way and we can check for the recommendations and we can ask ah client that is okay to move the vm manually so in a critical environment we will use manual let me give you a question critical environment okay i have one critical sql server okay critical sql server running on 2012 okay customer is customer is asking for cluster so what i did i have created two vms two virtual machines okay vm1 vm2 and you can configure windows failover cluster on top of it you know how to configure windows failover cluster sql is installed on top of it now this virtual machine both the virtual machines are running on one esxa host possible or not yes okay now due to h a or maybe h a is configured sorry my bad has configured due to some problem host went offline host went offline now your primary database server and your secondary database server both are down and what is the point of creating a cluster installing it on a cluster and confirming to the customer saying you are your database server is zero downtime now because you have two servers running in the back end but actually those two servers are running on one host the host got down then how we will handle this kind of situation okay the only the only question customer will ask why i should pay for two servers if both the servers are going on offline at a time which shouldn't happen for that reason only i am paying for two servers otherwise i am fine with one server you gave me two servers every month paying double bill and now you are saying that both the vms are down at a time how it will happen this is the problem what is the solution think of it tomorrow we will discuss about how to resolve these kind of issues tomorrow you got the query and the problem yeah i got it okay think of it we will discuss tomorrow let me stop the recording here we'll catch up tomorrow again same time

Channel: Srinivas Maganti

Views: 909

Rating: 4.6923075 out of 5

Keywords: vmware, esxi, vcenter 6.5, vcenter 6.7, vmware 6.7, vmware 6.5, esxi 6.7, vmware in telugu, vmware training, how to install vcenter server, azure, windows 2019, windows 2008 EOL, whats new in vmware, vmware vCenter appliance, vmware basics, esxi troubleshooting, postgressql, vmware to azure migration, vmware converter, devops, puppet, terraform, fibre channel storage, fc storage, vSAN storage, vmware storage, vmware vsphere 7.0, vmware with Kubenetes, vmware DRS, vsphere DRS

Id: bpQXWzTtz6c

Length: 44min 15sec (2655 seconds)

Published: Sat Apr 25 2020