• This topic has 8 replies, 4 voices, and was last updated February 28, 2017 by Harry S.

Performance issue on file server

  • Hello,

    we are using Zerto with an ESX-Host which is running a Windows file server, nothing else. When we move rather big files or directories (> 3GB) we get an deadlock situation. Could this be an issue of the data replication with Zerto. After about 30 seconds everything is performing in an normal manner.

    Kind regards.

    Dirk

    Have you tried removing protection from that system and or powering off the VRA on the host to see if your problem goes away?  Have you had this system protected for some time now and only now are seeing the issue? I don’t see how Zerto is causing this issue….

    Hi Brian,

    good idea, i’ll give it a try. It#s a brand new setup W2012R2 Server dealing as file server in a domain.

    For my understanding …. if i move a large amount of data on the server, it must concern the VRA Replication anyway?

    cu

    Hi Dirk,

    Sorry you are experiencing issues. What are you referring to when you say deadlock? Do you mean the VM? The host? ZVR behavior such as RPO? Other VMs?

    Thanks,

    Harry

    Follow me: www.twitter.com/HarrySiii

    Hi Harry,

    the fileserver is the only vm running on this ESX Host. Deadlog means, that the users will hang while brwosing in the explorer on a mapped drive from this file server for about 20 seconds.

    cu

    never seen an issue when replicating a VM and we replicate  a lot different kind of workloads and about 110 VMs no performance related issues

    Very odd and Carlos is correct, ZVR does not interfere with the protected VM at all, but it could be something else environmental. You could remove protection and even uninstall the VRA from that host and verify behavior.

    Where is your target site and storage? Are the source and target sharing the same infrastructure (hosts, clusters, storage, networks, etc)? If so, while ZVR will not impact the protected VMs, if it is heavily writing to the same storage also being read by this file server for its production data, I could see how you could receive performance impact, for example.

    Thanks,

    Harry

    Follow me: www.twitter.com/HarrySiii

    Thx a lot for your help.

    I’ll give it a try and deactivate Zerto. In fact the issue is easy to provoke ;-).

    Both ESX hosts use their local storage. The target host is about 15 miles away, connected via 100MBit.

    cu

    Dirk

    Dirk,

    Sounds good.  I can find and point you to recorded webinar that covers how this works at a high level, if that would help. Otherwise, you could reach out to support to troubleshoot this with you (or help you remove ZVR from the equation for that particular host/VM). You can also reach out to me (my first name and domain to email me) or your partner/account team to cover some of this architecture in more detail to better understand what ZVR is doing, and what your environment looks like.

    ~Harry

    Follow me: www.twitter.com/HarrySiii
You must be logged in to create new topics. Click here to login