This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Agentless Snapshot Question

Rapid Recovery 6.2.1...esxi 5.5...if i decide to do agentless on some of my servers, what is the method of the snapshot being created? I know the agent uses VSS and the local RR logs to know what to backup in each snapshot in windows but wasnt sure how Rapid Recovery knwos what has changed when backing up via VMWare agentless. I did notice in the setup of the core a mention about merging vmware snapshots....so i wasnt sure if using snapshots was how the agentless backup worked with VMWare

Parents
  • Agent-less backups for VMware are fairly straight forward, however they do have a few 'new' pieces to introduce if you're not familiar with the process. 

    The snapshot process is done on a .vmdk level within VMware, the snapshot is taken to remove the lock on the .vmdk and then the backup of (or from) the .vmdk is done. The changed blocks are tracking through VMware via Change Block Tracking (CBT) which is a VMware feature (https://kb.vmware.com/s/article/1020128). 

    You mention that you are on ESXi 5.5. The version should be adequate, however you'll want to make sure you're on recent build of 5.5. If you're running 5.5 from 2014 you might run into more snapshot problems and/or VMware conflicts than if you were running a build from late 2018. 

    The rise and fall of most agent-less DP solutions revolves around a few factors: 

    1. ESXi builds being up to date
    2. Virtual Center resources (RAM) - specifically, running the minimum resources needed to just install the VC isn't the best way to perform agent-less backups, 16 GB is a good start. 
    3. Job contention - agent-less backups alleviate resources on the RR core (generally) however they do increase overhead within ESXi. Just don't increase concurrent jobs to a bunch as the OS isn't be hit as hard

    A few things to keep in mind/think of regarding agent-less backups: 

    support.quest.com/.../backing-up-virtual-center-with-rapid-recovery

    support.quest.com/.../rapid-recovery-lan-free-san

    support.quest.com/.../rapid-recovery-esxi-6-6-5-6-7-permissions

Reply
  • Agent-less backups for VMware are fairly straight forward, however they do have a few 'new' pieces to introduce if you're not familiar with the process. 

    The snapshot process is done on a .vmdk level within VMware, the snapshot is taken to remove the lock on the .vmdk and then the backup of (or from) the .vmdk is done. The changed blocks are tracking through VMware via Change Block Tracking (CBT) which is a VMware feature (https://kb.vmware.com/s/article/1020128). 

    You mention that you are on ESXi 5.5. The version should be adequate, however you'll want to make sure you're on recent build of 5.5. If you're running 5.5 from 2014 you might run into more snapshot problems and/or VMware conflicts than if you were running a build from late 2018. 

    The rise and fall of most agent-less DP solutions revolves around a few factors: 

    1. ESXi builds being up to date
    2. Virtual Center resources (RAM) - specifically, running the minimum resources needed to just install the VC isn't the best way to perform agent-less backups, 16 GB is a good start. 
    3. Job contention - agent-less backups alleviate resources on the RR core (generally) however they do increase overhead within ESXi. Just don't increase concurrent jobs to a bunch as the OS isn't be hit as hard

    A few things to keep in mind/think of regarding agent-less backups: 

    support.quest.com/.../backing-up-virtual-center-with-rapid-recovery

    support.quest.com/.../rapid-recovery-lan-free-san

    support.quest.com/.../rapid-recovery-esxi-6-6-5-6-7-permissions

Children
  • Thanks for the info...it still isnt clear if the "it snapshot itsself "refers to creating a snapshot similar to using the snapshot manager in the VM console and manually creating a snapshot...where disk resources are used up on the shared drive array my vm's are using for disk storage....

    i have some work to do if i want to use this....CBT is not enabled on the VM's i have checked...granted i would probably do agent for windows...but would like to use agentless for my vm running my ip phone system and powerchute admin

  • These are VMware level snapshots yes. They would appear in snapshot manager, as they are on the host level yes (snapshots of the .vmdk files themselves). The snapshots again only contain the changes made while open, so you do need extra free space, but not double your used space or anything like that. 

    CBT get enabled through RR when you add teh VM to protection. Most DP solutions will do this for you, as we do. Since otherwise you'd have to be good at command line, or turn off the VM to do so. 

  • Is CBT only enabled for one VM at a time when that VM is protected in RR? The VMware doc mentions to remove snapshots before enabling CBT on a VM...i dont want to point RR at an esxi host and have every vm on that host enable cbt whether or not i am protecting agentless or not

  • It is per VM. CBT is VM based not host based. So if you have a host with 100 VMs and choose to backup 4 of them, just those 4 get CBT enabled.

  • In my test environment....running 6.1.300 core....trying to protect esxi 5.5....when i choose one server on the esxi host, i get a window (warning) with a link to a vmware article about enabling CBT...not sure if RR automatically enabling this is only available in newer ESXI versions??? it appears i have to do it manually in the VM config...

  • It does it for you automatically. The waning is just pointing you toward the VMware KB article to describe what CBT actually is (which includes the manual instructions). Which I believe is the same KB I referenced, unless either us or them changed one of our links. 

  • thanks...just did a test to protect a VM...worked as you stated...there is one vdmk disk with 2 partitions...i assume it creates a snapshot for each partition when RR does its backup? I also removed the machine from protection but noticed the ctk file is still there...and noticed the settings still set to true in config....does it hurt to leave CBT on after the machine is removed from protection?

  • Correct, on all accounts. CBT will stay, I have yet in my professional career seen a product that will go back an turn off CBT automatically, some have toggled to turn it off through their GUI, however not auto. There is zero harm in leaving it on. In fact it is amazing how many times people bring VMs into protection and they don't get notified to turn on CBT, as it had already been enabled from another 3rd party product over time.