Better if you have two ESXi hosts (HA and DRS not mandatory). Yes it install Splitter VIB on each ESXi. RP4VM comes with uninstaller tool, this tool remove everything like 1. RP4VM VM itself, 2. Splitter VIBs installed 3. RP4VM Plugin etc.
Yes, I find out myself I really need 2 ESXi hosts with vRPA-s. Because otherwise there is no any possibility to restore while vRPA lost and also without vRPA it even dont find datastores. But the uninstaller dont work at all - it shows "You cannot uninstall a cluster if only one cluster exists."......This sounds very stupid message and probably its bug. Or maybe it want I must first shut down vRPA VM? .....But I have now other problem. I installed first it with only one vRPA and now I want to incluse other. But I get error message "Incompatible VM VLAN configuration". This is also very stupid message. I have set up data network perfectly. Maybe it dont like I have one vRPA in distributed switch and other vRPA in standard switch? The information is just little small. Maybe I must converst also other ESXi to distributed switch.
Splitters and io-filter can be uninstalled by SSH and vCenter plugin can be uninstalled by vCenter MOB. Also in datastore it makes folder that must remove and also delete some vib-s from ESXi using WinSCP. But second node I cant install. I also tried installing cluster from command line, but I stuck - it asks some iSCSI port number. I have no idea what it is. I have no iSCSI for it, I have NFS instead for repository. But maybe I dont need this cluster (2xvRPA), I can install without problem when I choose only one vRPA node. Because I plan to replicate vCenter server. Example, when I have vRPA cluster, then when vCenter goes down, I anyway cant restore it with RecoverPoint, because its management can be accessed only by vCenter itself. I can only use replica and boot it up. Not able to use rollback or failback functionality when vCenter is down. And then I have no help from vRPA cluster also.
Have you read the installation instructions? Many of your questions are described there. And I suppose the problems you had were due to a misconfiguration.
I already readed this and there must not be any misconfiguration. I already tried all kind of possibilities. I think the software is badly developed. Also, those installation instructions are superficials. Example there wasnt at all talked about NFS. Because iSCSI dont allow to open datastore in both ESXi-s.
It's not talked about NFS because it is not necessary.
RecoverPoint for VMs replicates VMs on any type of storage that VMware supports including VMFS, NFS, vSAN, and VVols. Ensure that all ESXi hosts in the cluster where the vRPAs reside share the datastore for the repository VMDK
(page 25 in the installation guide)
That's all you need to know. It relies on the already connected datastores. Well, there's nothing to do here.
And it's only for the following reason that iSCSI will be discussed in more detail:
RecoverPoint for VMs supports TCP/IP or iSCSI communication modes between the splitters and the vRPAs. IP splitter communication mode is preferred.
(page 22 in the installation guide)
So, this has nothing to do with the datastore or storage connectivity.
>>It's not talked about NFS because it is not necessary.
Yes it is important. vRPA cluster need shared storage for repository. It even dont start installation without both ESXi-s see the same datastore (I tested it). Also, from documentation I readed it must be datastore and not shared virtual disk.
iSCSI I can let both ESXi initiator connect to the same iSCSI target, but vCenter dont allow me to mount datastore in this LUN in both ESXi simultaneously. When datastore is in one ESXi, then in other ESXi vCenter dont allow to mount the same datastore. There is even no option for that. Only option is to make new datastore and the window is empty for that iSCSI target.
>>>>RecoverPoint for VMs supports TCP/IP or iSCSI communication modes between the splitters and the vRPAs. IP splitter communication mode is preferred.
Splitter is vim and io-filter installed automaticallt into hypervisor while cluster install process. Communication must done between vmkernel adapter and vRPA vm (second virtual nic). Of course its TCP/IP, its virtual network. What iSCSI? What is iSCSI to do between vmkernel adapter and virtual machine virtual nic? But installer automatically installs this vRPA VM second virtual nic while install process. First after OVA installation, vRPA comes only with one virtual nic. I made this install process already successfully one time, but with only one vRPA node. Now I want to install 2 vRPA node cluster (that is recommended). But cluster installer even dont allow me to choose second ESXi. I have 2 ESXi, ESXI_12.40 and ESXi_12.42, where vRPA-s are installed. When I boot up vRPA in ESXI_12.40, open web management and start installing cluster, it shown in ESXi-s choice only ESXI_12.40. And ESXI_12.42 grayed out and error says "Incompatible VM VLAN configuration". When I boot up vRPA in ESXI_12.42, boot it up, start installing cluster, it shows I can choose only ESXI_12.42 and for ESXI_12.40 it shows "Incompatible VM VLAN configuration".
I have used the Recoverypoint solution years ago and have never had bigger problem with it. Personally, I also find the installation instructions clear enough and could install all components easily.
Of course its TCP/IP, its virtual network. What iSCSI? What is iSCSI to do between vmkernel adapter and virtual machine virtual nic?
Why do you think iSCSI can be configured on vmkernel level? And a vNIC doesn't care what is in the packets. So of course, iSCSI can be used for communication. A storage doesn't work any different.
Of course iSCSI need vmkernel adapter in "iSCSI Storage Adapter"-->"Network Port Binding". But question isne about iSCSI, it dont need iSCSI, I dont have any iSCSI for RecoveryPoint, I have NFS. But it needs, and this is written in documentation - vmkernel adapter for communication between splitter and vRPA VM (called data communication). There first is only one vnic, data vnic is installed while cluster installation. So, before claster install, installer dont care about data communication at all. It only cares about cluster communication between vRPA-1, vRPA-2, ESXi-1, ESXi-2 and vCenter. And this network is made, the traffic can reach each-other. I really dont understan what at all means error message "Incompatible VM VLAN configuration". VLAN is not related at all to VM-s, but instead portgroups. I have local virtual switches there. There is no any restrictions from VLAN side. Or they mean VLAN not as VLAN, but some virtual network, just communication between components. But there is communication. This is GUI based installer. Command-line installer talks about "iSCSI port number". I still dont understand what I must put there. What number? Where I get the number. Do it mean vNIC port number? To what NIC? Cluster communication or data communication? And in local vswitch there is not at all port numbers.
RecoverPoint error on creating destination journal. No matter - local or remote cluster, local or remote storage, automatic or manual journal cretation attempt. Just dont create journal for copy. And error is: "Failed to provision one or more journal volumes. Failure messages are: [createVirtualDiskViaMPIToStorageProcess Failed to complete successfully. ended with eCFVMDK_generalError]"
This isn't an EMC support forum. If you're having issues with one of their products please post on their forums.
I start using Zerto, instead of RecoverPoint. All work, no any problems.